Yong Wu
commited on
Commit
·
c72f96b
1
Parent(s):
d5ff5a1
Add new fused moe workloads for bs=1,16,64
Browse files- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048.jsonl +0 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_0565b5a8359c481388b71ecf62973e55.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_0a368dcac4954709ab9d0fdbd28e3106.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_15c1362ac77a4e4096e829900cc63697.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_19f4253c55df4ef6954d4ff2c3ffe62e.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_24f705bfd75a4690b702840ee97c493b.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_25ff572f80af4921acd3d243d70976b6.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_39af7c09dfff4dab81192aab73e2df4a.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_3f8d9099a78944eda656cf925180ec11.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4533832705a64b3fa27925aa8025214b.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4722c79f4cff460abcfab0f8a68a26ec.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_488901dae68a477386d73f623f57efb8.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4c1adfe1426048929e19cbc1088613a0.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_53ab07fa359e4414a49342b359bcc1f1.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_562b8e56207a48c894a67295b16590c0.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_599fba604ace449298d6d94db334a965.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_5ca00199d0ce46bdbc014c4666a89820.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_660f8dba8f3c4dda8c114294b42b003f.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_68baa0120cc548999c475164a85e0cc9.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_72bc329881824090b6ebfff6a11ffe35.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7453d3afb7ff4a2eabc780a7536277fd.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_754c17a313cf44389c607c279e7e3e1d.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7c6557cf9f214d44ad220f8285bb20ce.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_8461e216ba484a029096c9032a497ed0.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_9165a620d020414b9ec7e7aab77d7dc3.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_a395885e736a4b33b6a4d0019fc484fe.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_a4261006345c423f964ae287f31b730b.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_a638ad09b590491cbd3efc496b9d8f93.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_aba303756a304a4bbcd28e8db335f885.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_adf59ac40e3d46048297772d9baf728e.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_af6613531de140ccbf0e736240b72c4e.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_ba33d29bd554424cbfbb973ecafe2a3e.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c06e8f78dcf1403daaba3e70adc089ec.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c8000300402647e0a5467fa39a7228c3.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c86c2ede5f044442a200ae71800993e9.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cb9361ae63a140d5a92785868b36cd86.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cf5fef0d911c4708b10256831bd7eaf0.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d18188eb33ac480f9c0421ad9e91eced.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d908b28608834776bd664cdfa97d6874.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_da61fdd4d66d4617a0d708ff8805b56b.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_dc03da2bcf3e4adeb8df5f65985d23b2.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_e258d9a8cd0040b8b76420f0dbbd0b15.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f08191822108428a9a041bc9154b243c.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f30554cfa522432880df7f6cdd549a52.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f4f31704fad0480997c3bef41e58efec.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f68708cad9fb4cee83342e84f28d4f33.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_fa46f8ffc7484cb9b71876985a8c2171.safetensors +3 -0
- fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_filtered.jsonl +25 -0
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_0565b5a8359c481388b71ecf62973e55.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2181318e0f46c146f21ac3e9218f0e67c668d79a63e5507ceb902340e29f3082
|
| 3 |
+
size 1413713192
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_0a368dcac4954709ab9d0fdbd28e3106.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7e41892e245bbc4d65a9a95e4122cf917748cc4a0f3a674cd03d2dd9c3fef0f
|
| 3 |
+
size 1410043816
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_15c1362ac77a4e4096e829900cc63697.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0185c33859c325a27d7b224d9433f43edc622b109bc6784288883594e0dee6bb
|
| 3 |
+
size 1410094312
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_19f4253c55df4ef6954d4ff2c3ffe62e.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:93701885d840c7acef9a08d5b3ab5efb2c2c6047089e3fe96241e2411c7c992b
|
| 3 |
+
size 1410035400
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_24f705bfd75a4690b702840ee97c493b.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7e3e0f81095151f1260a6b59c29b33b0c7fd8830009d28cda67045350d2cc039
|
| 3 |
+
size 1409639840
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_25ff572f80af4921acd3d243d70976b6.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3e3943d06afeccd2e32540541d19317d55f1133dd7cfae49ff14ba416e2cb83c
|
| 3 |
+
size 1409782920
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_39af7c09dfff4dab81192aab73e2df4a.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e6c413eda3478332d265a13d58ea4a8ca71116a60ba84b07b986d5be78a083e0
|
| 3 |
+
size 1410052232
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_3f8d9099a78944eda656cf925180ec11.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:80dd9341a4d9c2958178c404abca4ea545d67b7cf753b918686f2e80713da6a6
|
| 3 |
+
size 1413713192
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4533832705a64b3fa27925aa8025214b.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bde83c5765cc10b0651a881f8ff3a40b46ba1b1e476364cafe35b3b6d2a18a26
|
| 3 |
+
size 1409665088
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4722c79f4cff460abcfab0f8a68a26ec.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e9f078c0548c70569eaf120b1e001c1779a52e61b2638c622724c0ee850b0e67
|
| 3 |
+
size 1409782920
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_488901dae68a477386d73f623f57efb8.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:90aa442c52241b89095ed08fa1c024bb13b693e0b0516d01432af0f6ca6bc9ac
|
| 3 |
+
size 1409766088
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4c1adfe1426048929e19cbc1088613a0.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e42d64e9f4c6cde16bbbf583a0eaa8550eb51a8c10e1a7e4aef19451a05fecc7
|
| 3 |
+
size 1410102728
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_53ab07fa359e4414a49342b359bcc1f1.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2b0146eceabe88b54b20e880cb3b79b2c34b555317dcd3032090110ec9c6081
|
| 3 |
+
size 1410111144
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_562b8e56207a48c894a67295b16590c0.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:26359ac43c85b1d4601bb0f2c57fb31a7520e5327eb7cc68caab386c29d06a03
|
| 3 |
+
size 1413654280
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_599fba604ace449298d6d94db334a965.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bb2f2d5dd1f402f527e03d16542af203d9aab6647790ed22bd22fa3a7ca351f2
|
| 3 |
+
size 1409690336
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_5ca00199d0ce46bdbc014c4666a89820.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9b385c32597fc4023ec178d10515436f997e879efb3972b5f7b640c10f57ab96
|
| 3 |
+
size 1510185816
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_660f8dba8f3c4dda8c114294b42b003f.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4442d50fd57ed33fa7d72ad7d03c5685207119cff5a4b65af1e49c7787cd2586
|
| 3 |
+
size 1410102728
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_68baa0120cc548999c475164a85e0cc9.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ddf0eecb7b538f39fec89abbb0d1fe93ec5811c5b2e2a638ca6c5e95b698df7a
|
| 3 |
+
size 1410153224
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_72bc329881824090b6ebfff6a11ffe35.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a25e4afbf8840fdd0a14902da097d72ff1d0d24a0bc761d50b4f81abf66d944d
|
| 3 |
+
size 1510185816
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7453d3afb7ff4a2eabc780a7536277fd.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:66c7daa9800b38bc1af4600d9479ee591e1ad4c5f78c2f4790f7087f543f2ebe
|
| 3 |
+
size 1409900744
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_754c17a313cf44389c607c279e7e3e1d.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9dcf45601776a0e5e2855d4ffc30b64cbc2fef32bdeaddcf4ff6a27b890ad1a9
|
| 3 |
+
size 1410085896
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7c6557cf9f214d44ad220f8285bb20ce.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1911d67fce513d2a3451e01d648a0a555909fd6f2d27f72c29bf81052ee14398
|
| 3 |
+
size 1410111144
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_8461e216ba484a029096c9032a497ed0.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c24a033f3c63c1137bcf33f2f7265bddbbda242e21e884b961217bd8f258e869
|
| 3 |
+
size 1409690336
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_9165a620d020414b9ec7e7aab77d7dc3.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:844ac5ee1fd66232bed2c7c9e268f0550143a251713d0c1ee4590555f2c13642
|
| 3 |
+
size 1410069064
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_a395885e736a4b33b6a4d0019fc484fe.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fdb9e8d36ed4190f688f5f67b5667dd77f19ae0886829395fbb0ca3007cef463
|
| 3 |
+
size 1409757672
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_a4261006345c423f964ae287f31b730b.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e321abdb217573a0820b8285045a09bb8461664340c7cc5c6d958012299902c2
|
| 3 |
+
size 1410085896
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_a638ad09b590491cbd3efc496b9d8f93.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0021e70adb057f797919a4e2cb492d247143b6dd5ec73dbd6bd18cf40cd37fcc
|
| 3 |
+
size 1410077480
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_aba303756a304a4bbcd28e8db335f885.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32243804898d57d7c7ce91f454d660c4381dfb06b0601ac1ac04581a8d0c111b
|
| 3 |
+
size 1410060648
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_adf59ac40e3d46048297772d9baf728e.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:29a455997a4a6b5ed44f24b8895049524a00f8a7d8a0ac050098542dff8746d4
|
| 3 |
+
size 1410153224
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_af6613531de140ccbf0e736240b72c4e.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:88293c21047fa260cb3c0bec4065647cc3f94b8183d7e201b1903b5599229d6b
|
| 3 |
+
size 1409665088
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_ba33d29bd554424cbfbb973ecafe2a3e.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:757a58cd9b1045ffb23d7dbffcc5f9f16b07b522b4e314ee1328f4df293261e2
|
| 3 |
+
size 1409900744
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c06e8f78dcf1403daaba3e70adc089ec.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b320ee1d8b596f3dc305397ac8e18f145ed229bb4e821036c04cc5aeb1dc2508
|
| 3 |
+
size 1410052232
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c8000300402647e0a5467fa39a7228c3.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1699cb5ec4667d15861c68043c3ad9e7866075b9f2c3bae89f87d39a1ab44ffd
|
| 3 |
+
size 1410060648
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c86c2ede5f044442a200ae71800993e9.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47a19c461b71bfdbefd224d6b8e7cb3aeea156059fa9917c4293b2503f2b2750
|
| 3 |
+
size 1409766088
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cb9361ae63a140d5a92785868b36cd86.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:012f63590c061a0fa02c544b9f9f0f453fa3e6fbb2b75ddc1a62d96307ff9edf
|
| 3 |
+
size 1410043816
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cf5fef0d911c4708b10256831bd7eaf0.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e7d9577366e26433e22693e32de4efb2a45190b5ad4c5e3609905ee3bfb1ef50
|
| 3 |
+
size 1409749256
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d18188eb33ac480f9c0421ad9e91eced.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3f753cd4cf43797c2dfe5a5640b61e02a32cb9846e863c45a672cb75f6aa0474
|
| 3 |
+
size 1410026984
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d908b28608834776bd664cdfa97d6874.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1a3e666595b758c5b86f54f627ac005dd8d184adcd0b28004307180f95f7c375
|
| 3 |
+
size 1410035400
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_da61fdd4d66d4617a0d708ff8805b56b.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7bed74209a112388a6a47d2cd33f4b5a6baeaa1b0712ee26ddf7a3936ef2a56e
|
| 3 |
+
size 1410077480
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_dc03da2bcf3e4adeb8df5f65985d23b2.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:84b3ea9cf94f63741834e6fbd3e35732fbcf018c2f8a7698fea218a4f4aeaab9
|
| 3 |
+
size 1409757672
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_e258d9a8cd0040b8b76420f0dbbd0b15.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab09fad20cab679d60c79b7e9cbef7fe8317e1e13b7cc01730a8d04b5f8cee12
|
| 3 |
+
size 1409639840
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f08191822108428a9a041bc9154b243c.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:48133335223c8601e5c89c2ed0cb49bfd444fb831bd542ea03488f1b7eb2fbb6
|
| 3 |
+
size 1410119560
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f30554cfa522432880df7f6cdd549a52.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:58e74d2c13d4a5b67c733962845bc92ec4a49655d8ae0a949a14eaf11bf95607
|
| 3 |
+
size 1409749256
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f4f31704fad0480997c3bef41e58efec.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dab3f37e2fdabd1539386ef8bf3a10fb010d936bf3fb6b6df30a3f9b01faf645
|
| 3 |
+
size 1528322296
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f68708cad9fb4cee83342e84f28d4f33.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:813adf812c68dfad4417ee2cb063bb7da1a6478781f3036820cc2ba847d1eed3
|
| 3 |
+
size 1410026984
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_fa46f8ffc7484cb9b71876985a8c2171.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c5cb25e952cd5d27f476caa171bbf91a2a488c310abb63d837787d08921331a
|
| 3 |
+
size 1413654280
|
fused_moe_wl/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_filtered.jsonl
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "70650a3b-0854-40fd-9bb5-56d1e83dab79", "axes": {"seq_len": 7}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_8461e216ba484a029096c9032a497ed0.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_8461e216ba484a029096c9032a497ed0.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_8461e216ba484a029096c9032a497ed0.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_8461e216ba484a029096c9032a497ed0.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_8461e216ba484a029096c9032a497ed0.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_8461e216ba484a029096c9032a497ed0.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_8461e216ba484a029096c9032a497ed0.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_8461e216ba484a029096c9032a497ed0.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 128}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 2 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "eb06a62f-23e3-4ee4-b4ee-3a4b2c23aca5", "axes": {"seq_len": 1}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_24f705bfd75a4690b702840ee97c493b.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_24f705bfd75a4690b702840ee97c493b.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_24f705bfd75a4690b702840ee97c493b.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_24f705bfd75a4690b702840ee97c493b.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_24f705bfd75a4690b702840ee97c493b.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_24f705bfd75a4690b702840ee97c493b.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_24f705bfd75a4690b702840ee97c493b.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_24f705bfd75a4690b702840ee97c493b.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 128}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 3 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "64d964f5-5887-4a2b-b542-1f33aa7c3ff1", "axes": {"seq_len": 32}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7453d3afb7ff4a2eabc780a7536277fd.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7453d3afb7ff4a2eabc780a7536277fd.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7453d3afb7ff4a2eabc780a7536277fd.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7453d3afb7ff4a2eabc780a7536277fd.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7453d3afb7ff4a2eabc780a7536277fd.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7453d3afb7ff4a2eabc780a7536277fd.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7453d3afb7ff4a2eabc780a7536277fd.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7453d3afb7ff4a2eabc780a7536277fd.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 0}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 4 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "da1b42f9-fc1f-448f-bba7-06154f8e71fa", "axes": {"seq_len": 18}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4722c79f4cff460abcfab0f8a68a26ec.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4722c79f4cff460abcfab0f8a68a26ec.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4722c79f4cff460abcfab0f8a68a26ec.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4722c79f4cff460abcfab0f8a68a26ec.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4722c79f4cff460abcfab0f8a68a26ec.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4722c79f4cff460abcfab0f8a68a26ec.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4722c79f4cff460abcfab0f8a68a26ec.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4722c79f4cff460abcfab0f8a68a26ec.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 0}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 5 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "9901ef73-0072-42bd-988a-ac11ad193735", "axes": {"seq_len": 478}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_fa46f8ffc7484cb9b71876985a8c2171.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_fa46f8ffc7484cb9b71876985a8c2171.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_fa46f8ffc7484cb9b71876985a8c2171.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_fa46f8ffc7484cb9b71876985a8c2171.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_fa46f8ffc7484cb9b71876985a8c2171.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_fa46f8ffc7484cb9b71876985a8c2171.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_fa46f8ffc7484cb9b71876985a8c2171.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_fa46f8ffc7484cb9b71876985a8c2171.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 192}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 6 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "e50042a3-9e76-4d78-a307-2dcf272cf600", "axes": {"seq_len": 485}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_3f8d9099a78944eda656cf925180ec11.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_3f8d9099a78944eda656cf925180ec11.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_3f8d9099a78944eda656cf925180ec11.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_3f8d9099a78944eda656cf925180ec11.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_3f8d9099a78944eda656cf925180ec11.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_3f8d9099a78944eda656cf925180ec11.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_3f8d9099a78944eda656cf925180ec11.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_3f8d9099a78944eda656cf925180ec11.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 192}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 7 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "806265ee-5e6b-4ed3-925c-d0b6a251e275", "axes": {"seq_len": 16}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c86c2ede5f044442a200ae71800993e9.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c86c2ede5f044442a200ae71800993e9.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c86c2ede5f044442a200ae71800993e9.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c86c2ede5f044442a200ae71800993e9.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c86c2ede5f044442a200ae71800993e9.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c86c2ede5f044442a200ae71800993e9.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c86c2ede5f044442a200ae71800993e9.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c86c2ede5f044442a200ae71800993e9.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 224}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 8 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "5333abde-5617-4802-9058-e945aac6cb54", "axes": {"seq_len": 15}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_dc03da2bcf3e4adeb8df5f65985d23b2.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_dc03da2bcf3e4adeb8df5f65985d23b2.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_dc03da2bcf3e4adeb8df5f65985d23b2.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_dc03da2bcf3e4adeb8df5f65985d23b2.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_dc03da2bcf3e4adeb8df5f65985d23b2.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_dc03da2bcf3e4adeb8df5f65985d23b2.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_dc03da2bcf3e4adeb8df5f65985d23b2.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_dc03da2bcf3e4adeb8df5f65985d23b2.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 96}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 9 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "1d7a6503-a0b5-4cad-b827-5195fcdb2b78", "axes": {"seq_len": 14}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f30554cfa522432880df7f6cdd549a52.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f30554cfa522432880df7f6cdd549a52.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f30554cfa522432880df7f6cdd549a52.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f30554cfa522432880df7f6cdd549a52.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f30554cfa522432880df7f6cdd549a52.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f30554cfa522432880df7f6cdd549a52.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f30554cfa522432880df7f6cdd549a52.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f30554cfa522432880df7f6cdd549a52.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 160}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 10 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "22b0bbf9-9705-4e1b-b0c9-bf011e4ccfd6", "axes": {"seq_len": 4}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_af6613531de140ccbf0e736240b72c4e.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_af6613531de140ccbf0e736240b72c4e.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_af6613531de140ccbf0e736240b72c4e.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_af6613531de140ccbf0e736240b72c4e.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_af6613531de140ccbf0e736240b72c4e.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_af6613531de140ccbf0e736240b72c4e.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_af6613531de140ccbf0e736240b72c4e.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_af6613531de140ccbf0e736240b72c4e.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 0}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 11 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "8e155616-684c-46af-bf44-7db87ea03bb8", "axes": {"seq_len": 14103}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f4f31704fad0480997c3bef41e58efec.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f4f31704fad0480997c3bef41e58efec.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f4f31704fad0480997c3bef41e58efec.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f4f31704fad0480997c3bef41e58efec.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f4f31704fad0480997c3bef41e58efec.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f4f31704fad0480997c3bef41e58efec.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f4f31704fad0480997c3bef41e58efec.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f4f31704fad0480997c3bef41e58efec.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 224}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 12 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "bdbcf40c-295b-4389-aebf-e89180483335", "axes": {"seq_len": 11948}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_72bc329881824090b6ebfff6a11ffe35.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_72bc329881824090b6ebfff6a11ffe35.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_72bc329881824090b6ebfff6a11ffe35.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_72bc329881824090b6ebfff6a11ffe35.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_72bc329881824090b6ebfff6a11ffe35.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_72bc329881824090b6ebfff6a11ffe35.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_72bc329881824090b6ebfff6a11ffe35.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_72bc329881824090b6ebfff6a11ffe35.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 224}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 13 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "318e0446-a189-4aad-a4bf-d7bae7e8ce36", "axes": {"seq_len": 62}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_adf59ac40e3d46048297772d9baf728e.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_adf59ac40e3d46048297772d9baf728e.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_adf59ac40e3d46048297772d9baf728e.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_adf59ac40e3d46048297772d9baf728e.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_adf59ac40e3d46048297772d9baf728e.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_adf59ac40e3d46048297772d9baf728e.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_adf59ac40e3d46048297772d9baf728e.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_adf59ac40e3d46048297772d9baf728e.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 0}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 14 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "8be0469a-018b-4c30-905f-7ef152fe81d3", "axes": {"seq_len": 58}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f08191822108428a9a041bc9154b243c.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f08191822108428a9a041bc9154b243c.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f08191822108428a9a041bc9154b243c.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f08191822108428a9a041bc9154b243c.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f08191822108428a9a041bc9154b243c.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f08191822108428a9a041bc9154b243c.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f08191822108428a9a041bc9154b243c.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f08191822108428a9a041bc9154b243c.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 64}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 15 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "f4ab9891-bd8a-450f-8c7a-c921d5f83762", "axes": {"seq_len": 57}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7c6557cf9f214d44ad220f8285bb20ce.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7c6557cf9f214d44ad220f8285bb20ce.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7c6557cf9f214d44ad220f8285bb20ce.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7c6557cf9f214d44ad220f8285bb20ce.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7c6557cf9f214d44ad220f8285bb20ce.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7c6557cf9f214d44ad220f8285bb20ce.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7c6557cf9f214d44ad220f8285bb20ce.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_7c6557cf9f214d44ad220f8285bb20ce.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 0}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 16 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "4cfb057f-18f5-4a71-9fce-1eaee691caab", "axes": {"seq_len": 56}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4c1adfe1426048929e19cbc1088613a0.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4c1adfe1426048929e19cbc1088613a0.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4c1adfe1426048929e19cbc1088613a0.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4c1adfe1426048929e19cbc1088613a0.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4c1adfe1426048929e19cbc1088613a0.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4c1adfe1426048929e19cbc1088613a0.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4c1adfe1426048929e19cbc1088613a0.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_4c1adfe1426048929e19cbc1088613a0.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 32}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 17 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "23624b64-1ee4-41d9-b57e-d1002a324d24", "axes": {"seq_len": 55}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_15c1362ac77a4e4096e829900cc63697.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_15c1362ac77a4e4096e829900cc63697.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_15c1362ac77a4e4096e829900cc63697.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_15c1362ac77a4e4096e829900cc63697.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_15c1362ac77a4e4096e829900cc63697.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_15c1362ac77a4e4096e829900cc63697.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_15c1362ac77a4e4096e829900cc63697.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_15c1362ac77a4e4096e829900cc63697.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 96}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 18 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "44ee3fc8-7363-4b6f-bdf9-89e8f6a95442", "axes": {"seq_len": 54}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_754c17a313cf44389c607c279e7e3e1d.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_754c17a313cf44389c607c279e7e3e1d.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_754c17a313cf44389c607c279e7e3e1d.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_754c17a313cf44389c607c279e7e3e1d.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_754c17a313cf44389c607c279e7e3e1d.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_754c17a313cf44389c607c279e7e3e1d.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_754c17a313cf44389c607c279e7e3e1d.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_754c17a313cf44389c607c279e7e3e1d.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 224}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 19 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "e4b12c0c-9234-406d-86ec-8042c21601a4", "axes": {"seq_len": 53}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_da61fdd4d66d4617a0d708ff8805b56b.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_da61fdd4d66d4617a0d708ff8805b56b.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_da61fdd4d66d4617a0d708ff8805b56b.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_da61fdd4d66d4617a0d708ff8805b56b.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_da61fdd4d66d4617a0d708ff8805b56b.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_da61fdd4d66d4617a0d708ff8805b56b.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_da61fdd4d66d4617a0d708ff8805b56b.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_da61fdd4d66d4617a0d708ff8805b56b.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 224}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 20 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "7b5b4b5f-4cf3-457b-980d-ce312a0009c9", "axes": {"seq_len": 52}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_9165a620d020414b9ec7e7aab77d7dc3.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_9165a620d020414b9ec7e7aab77d7dc3.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_9165a620d020414b9ec7e7aab77d7dc3.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_9165a620d020414b9ec7e7aab77d7dc3.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_9165a620d020414b9ec7e7aab77d7dc3.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_9165a620d020414b9ec7e7aab77d7dc3.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_9165a620d020414b9ec7e7aab77d7dc3.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_9165a620d020414b9ec7e7aab77d7dc3.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 0}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 21 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "107d90dc-ac3b-4e8c-b571-1f4f5fb9391e", "axes": {"seq_len": 51}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c8000300402647e0a5467fa39a7228c3.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c8000300402647e0a5467fa39a7228c3.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c8000300402647e0a5467fa39a7228c3.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c8000300402647e0a5467fa39a7228c3.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c8000300402647e0a5467fa39a7228c3.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c8000300402647e0a5467fa39a7228c3.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c8000300402647e0a5467fa39a7228c3.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c8000300402647e0a5467fa39a7228c3.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 96}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 22 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "4581ee01-b27a-405f-ab93-75b97783f5d8", "axes": {"seq_len": 50}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c06e8f78dcf1403daaba3e70adc089ec.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c06e8f78dcf1403daaba3e70adc089ec.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c06e8f78dcf1403daaba3e70adc089ec.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c06e8f78dcf1403daaba3e70adc089ec.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c06e8f78dcf1403daaba3e70adc089ec.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c06e8f78dcf1403daaba3e70adc089ec.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c06e8f78dcf1403daaba3e70adc089ec.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_c06e8f78dcf1403daaba3e70adc089ec.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 64}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 23 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "73958681-aee9-4aa5-b911-9c89f0127236", "axes": {"seq_len": 49}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cb9361ae63a140d5a92785868b36cd86.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cb9361ae63a140d5a92785868b36cd86.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cb9361ae63a140d5a92785868b36cd86.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cb9361ae63a140d5a92785868b36cd86.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cb9361ae63a140d5a92785868b36cd86.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cb9361ae63a140d5a92785868b36cd86.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cb9361ae63a140d5a92785868b36cd86.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_cb9361ae63a140d5a92785868b36cd86.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 32}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 24 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "c5341121-9604-4ba6-908a-ea7731b209bd", "axes": {"seq_len": 48}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d908b28608834776bd664cdfa97d6874.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d908b28608834776bd664cdfa97d6874.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d908b28608834776bd664cdfa97d6874.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d908b28608834776bd664cdfa97d6874.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d908b28608834776bd664cdfa97d6874.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d908b28608834776bd664cdfa97d6874.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d908b28608834776bd664cdfa97d6874.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_d908b28608834776bd664cdfa97d6874.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 96}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|
| 25 |
+
{"definition": "moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048", "solution": null, "workload": {"uuid": "f00c7d3a-70e8-4e86-94c2-6d81f798faae", "axes": {"seq_len": 47}, "inputs": {"routing_logits": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f68708cad9fb4cee83342e84f28d4f33.safetensors", "tensor_key": "routing_logits"}, "hidden_states": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f68708cad9fb4cee83342e84f28d4f33.safetensors", "tensor_key": "hidden_states"}, "hidden_states_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f68708cad9fb4cee83342e84f28d4f33.safetensors", "tensor_key": "hidden_states_scale"}, "routing_bias": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f68708cad9fb4cee83342e84f28d4f33.safetensors", "tensor_key": "routing_bias"}, "gemm1_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f68708cad9fb4cee83342e84f28d4f33.safetensors", "tensor_key": "gemm1_weights"}, "gemm1_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f68708cad9fb4cee83342e84f28d4f33.safetensors", "tensor_key": "gemm1_weights_scale"}, "gemm2_weights": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f68708cad9fb4cee83342e84f28d4f33.safetensors", "tensor_key": "gemm2_weights"}, "gemm2_weights_scale": {"type": "safetensors", "path": "./blob/workloads/moe/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048/moe_fp8_block_scale_ds_routing_topk8_ng8_kg4_e32_h7168_i2048_f68708cad9fb4cee83342e84f28d4f33.safetensors", "tensor_key": "gemm2_weights_scale"}, "local_expert_offset": {"type": "scalar", "value": 224}, "routed_scaling_factor": {"type": "scalar", "value": 2.5}}}, "evaluation": null}
|