Hugging Face's logo

optimum-internal-testing
/
neuron-testing-cache

Model card Files Community
dacorvo HF Staff commited on
Commit
e315806
·
verified ·
1 Parent(s): e852b17

Synchronizing local compiler cache.

Browse files
Files changed (29) hide show
  1. .gitattributes +8 -0
  2. neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.3.0.dev2/qwen3/Qwen/Qwen3-32B/a10939c94ae8684f8a5d.json +72 -0
  3. neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.3.0.dev3/qwen3/Qwen/Qwen3-0.6B/8adeca2486f370c9e74a.json +72 -0
  4. neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/compile_flags.json +1 -0
  5. neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/model.done +0 -0
  6. neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/model.hlo_module.pb +3 -0
  7. neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/model.neff +3 -0
  8. neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/wrapped_neff.hlo +3 -0
  9. neuronxcc-2.17.194.0+d312836f/MODULE_604057e5430f1ff4adef+431f5505/compile_flags.json +1 -0
  10. neuronxcc-2.17.194.0+d312836f/MODULE_604057e5430f1ff4adef+431f5505/model.done +0 -0
  11. neuronxcc-2.17.194.0+d312836f/MODULE_604057e5430f1ff4adef+431f5505/model.hlo_module.pb +3 -0
  12. neuronxcc-2.17.194.0+d312836f/MODULE_604057e5430f1ff4adef+431f5505/model.neff +3 -0
  13. neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/compile_flags.json +1 -0
  14. neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/model.done +0 -0
  15. neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/model.hlo_module.pb +3 -0
  16. neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/model.neff +3 -0
  17. neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/wrapped_neff.hlo +3 -0
  18. neuronxcc-2.17.194.0+d312836f/MODULE_96d97f202e238feb2c21+bfe5714b/compile_flags.json +1 -0
  19. neuronxcc-2.17.194.0+d312836f/MODULE_96d97f202e238feb2c21+bfe5714b/model.done +0 -0
  20. neuronxcc-2.17.194.0+d312836f/MODULE_96d97f202e238feb2c21+bfe5714b/model.hlo_module.pb +3 -0
  21. neuronxcc-2.17.194.0+d312836f/MODULE_96d97f202e238feb2c21+bfe5714b/model.neff +3 -0
  22. neuronxcc-2.17.194.0+d312836f/MODULE_9bce7228204bd2d4577b+431f5505/compile_flags.json +1 -0
  23. neuronxcc-2.17.194.0+d312836f/MODULE_9bce7228204bd2d4577b+431f5505/model.done +0 -0
  24. neuronxcc-2.17.194.0+d312836f/MODULE_9bce7228204bd2d4577b+431f5505/model.hlo_module.pb +3 -0
  25. neuronxcc-2.17.194.0+d312836f/MODULE_9bce7228204bd2d4577b+431f5505/model.neff +3 -0
  26. neuronxcc-2.17.194.0+d312836f/MODULE_9cd10f645f9c9838f49a+bfe5714b/compile_flags.json +1 -0
  27. neuronxcc-2.17.194.0+d312836f/MODULE_9cd10f645f9c9838f49a+bfe5714b/model.done +0 -0
  28. neuronxcc-2.17.194.0+d312836f/MODULE_9cd10f645f9c9838f49a+bfe5714b/model.hlo_module.pb +3 -0
  29. neuronxcc-2.17.194.0+d312836f/MODULE_9cd10f645f9c9838f49a+bfe5714b/model.neff +3 -0
.gitattributes CHANGED
@@ -3349,3 +3349,11 @@ neuronxcc-2.17.194.0+d312836f/MODULE_ae5ca4b91afd03b04b25+63ea7cf2/wrapped_neff.
3349
  neuronxcc-2.17.194.0+d312836f/MODULE_48fa12ade411c141306d+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
3350
  neuronxcc-2.17.194.0+d312836f/MODULE_6526b7e63165e490d82d+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
3351
  neuronxcc-2.17.194.0+d312836f/MODULE_6526b7e63165e490d82d+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
3349
  neuronxcc-2.17.194.0+d312836f/MODULE_48fa12ade411c141306d+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
3350
  neuronxcc-2.17.194.0+d312836f/MODULE_6526b7e63165e490d82d+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
3351
  neuronxcc-2.17.194.0+d312836f/MODULE_6526b7e63165e490d82d+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
3352
+ neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
3353
+ neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
3354
+ neuronxcc-2.17.194.0+d312836f/MODULE_604057e5430f1ff4adef+431f5505/model.neff filter=lfs diff=lfs merge=lfs -text
3355
+ neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
3356
+ neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
3357
+ neuronxcc-2.17.194.0+d312836f/MODULE_96d97f202e238feb2c21+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
3358
+ neuronxcc-2.17.194.0+d312836f/MODULE_9bce7228204bd2d4577b+431f5505/model.neff filter=lfs diff=lfs merge=lfs -text
3359
+ neuronxcc-2.17.194.0+d312836f/MODULE_9cd10f645f9c9838f49a+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.3.0.dev2/qwen3/Qwen/Qwen3-32B/a10939c94ae8684f8a5d.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen3-32B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "head_dim": 128,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 5120,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 25600,
15
+ "max_position_embeddings": 40960,
16
+ "max_window_layers": 64,
17
+ "model_type": "qwen3",
18
+ "neuron": {
19
+ "_serialized_key": "NxDNeuronConfig",
20
+ "async_mode": false,
21
+ "attn_kernel_enabled": false,
22
+ "batch_size": 1,
23
+ "capacity_factor": null,
24
+ "cc_pipeline_tiling_factor": 2,
25
+ "checkpoint_id": "Qwen/Qwen3-32B",
26
+ "checkpoint_revision": "d47b0d4ae4b48fde975756bf360a63a9cca8d470",
27
+ "continuous_batching": false,
28
+ "enable_bucketing": false,
29
+ "ep_degree": 1,
30
+ "flash_decoding_enabled": false,
31
+ "fused_qkv": true,
32
+ "glu_mlp": true,
33
+ "is_chunked_prefill": false,
34
+ "local_ranks_size": 8,
35
+ "logical_nc_config": 1,
36
+ "max_batch_size": 1,
37
+ "max_context_length": 4096,
38
+ "max_topk": 256,
39
+ "mlp_kernel_enabled": false,
40
+ "mlp_kernel_fuse_residual_add": false,
41
+ "n_active_tokens": 4096,
42
+ "neuronxcc_version": "2.17.194.0+d312836f",
43
+ "num_cores_per_group": 1,
44
+ "on_device_sampling": true,
45
+ "optimum_neuron_version": "0.3.0.dev2",
46
+ "output_logits": false,
47
+ "padding_side": "right",
48
+ "pp_degree": 1,
49
+ "qk_layernorm": false,
50
+ "qkv_kernel_enabled": false,
51
+ "rpl_reduce_dtype": "bfloat16",
52
+ "sequence_length": 4096,
53
+ "sequence_parallel_enabled": false,
54
+ "speculation_length": 0,
55
+ "start_rank_id": 0,
56
+ "target": null,
57
+ "torch_dtype": "bfloat16",
58
+ "tp_degree": 8,
59
+ "vocab_parallel": false
60
+ },
61
+ "num_attention_heads": 64,
62
+ "num_hidden_layers": 64,
63
+ "num_key_value_heads": 8,
64
+ "rms_norm_eps": 1e-06,
65
+ "rope_scaling": null,
66
+ "rope_theta": 1000000,
67
+ "sliding_window": null,
68
+ "tie_word_embeddings": false,
69
+ "use_cache": true,
70
+ "use_sliding_window": false,
71
+ "vocab_size": 151936
72
+ }
neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.3.0.dev3/qwen3/Qwen/Qwen3-0.6B/8adeca2486f370c9e74a.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen3-0.6B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "head_dim": 128,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 1024,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 3072,
15
+ "max_position_embeddings": 40960,
16
+ "max_window_layers": 28,
17
+ "model_type": "qwen3",
18
+ "neuron": {
19
+ "_serialized_key": "NxDNeuronConfig",
20
+ "async_mode": false,
21
+ "attn_kernel_enabled": false,
22
+ "batch_size": 4,
23
+ "capacity_factor": null,
24
+ "cc_pipeline_tiling_factor": 2,
25
+ "checkpoint_id": "Qwen/Qwen3-0.6B",
26
+ "checkpoint_revision": "e6de91484c29aa9480d55605af694f39b081c455",
27
+ "continuous_batching": true,
28
+ "enable_bucketing": false,
29
+ "ep_degree": 1,
30
+ "flash_decoding_enabled": false,
31
+ "fused_qkv": true,
32
+ "glu_mlp": true,
33
+ "is_chunked_prefill": false,
34
+ "local_ranks_size": 2,
35
+ "logical_nc_config": 1,
36
+ "max_batch_size": 4,
37
+ "max_context_length": 4096,
38
+ "max_topk": 256,
39
+ "mlp_kernel_enabled": false,
40
+ "mlp_kernel_fuse_residual_add": false,
41
+ "n_active_tokens": 4096,
42
+ "neuronxcc_version": "2.17.194.0+d312836f",
43
+ "num_cores_per_group": 1,
44
+ "on_device_sampling": false,
45
+ "optimum_neuron_version": "0.3.0.dev3",
46
+ "output_logits": false,
47
+ "padding_side": "right",
48
+ "pp_degree": 1,
49
+ "qk_layernorm": false,
50
+ "qkv_kernel_enabled": false,
51
+ "rpl_reduce_dtype": "bfloat16",
52
+ "sequence_length": 4096,
53
+ "sequence_parallel_enabled": false,
54
+ "speculation_length": 0,
55
+ "start_rank_id": 0,
56
+ "target": null,
57
+ "torch_dtype": "bfloat16",
58
+ "tp_degree": 2,
59
+ "vocab_parallel": false
60
+ },
61
+ "num_attention_heads": 16,
62
+ "num_hidden_layers": 28,
63
+ "num_key_value_heads": 8,
64
+ "rms_norm_eps": 1e-06,
65
+ "rope_scaling": null,
66
+ "rope_theta": 1000000,
67
+ "sliding_window": null,
68
+ "tie_word_embeddings": true,
69
+ "use_cache": true,
70
+ "use_sliding_window": false,
71
+ "vocab_size": 151936
72
+ }
neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ "--auto-cast=none --model-type=transformer --tensorizer-options='--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-dge-dma --vectorize-strided-dma ' -O2 --internal-num-neuroncores-per-sengine=1 --logfile=/tmp/nxd_model/token_generation_model/_tp0_bk0/log-neuron-cc.txt --enable-internal-neff-wrapper"
neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa0b2112e492d4b7381ad82fdf6ae9d3f3b53d1ab7823bcf6a8f530b35fc6e04
3
+ size 694657
neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e24141020ae7acce088080a4c4fd98075257c4d745829c5e2ebe4bee931e296a
3
+ size 2601984
neuronxcc-2.17.194.0+d312836f/MODULE_23d27f640534fb2c4304+165e9558/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:84c006a9d07ecb1848cc408d972415f2f87454be86b3c9c2385bcf0116cf5dcd
3
+ size 2732414
neuronxcc-2.17.194.0+d312836f/MODULE_604057e5430f1ff4adef+431f5505/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ "--model-type=transformer -O1 --lnc=1 --internal-hlo2tensorizer-options=--experimental-unsafe-fp8e4m3fn-as-fp8e4m3 --logfile=/tmp/nxd_model/layout_opt/log-neuron-cc.txt"
neuronxcc-2.17.194.0+d312836f/MODULE_604057e5430f1ff4adef+431f5505/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_604057e5430f1ff4adef+431f5505/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:10998616509ace73dd7592c65bcc632cea224f9d4fbeab8158d7c2a6a9d63949
3
+ size 298017
neuronxcc-2.17.194.0+d312836f/MODULE_604057e5430f1ff4adef+431f5505/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f24ccfafeda537f2cb70b15c1a68065605cefa5c3db1664f05303561d292ae2
3
+ size 7128064
neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ "--auto-cast=none --model-type=transformer --tensorizer-options='--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-dge-dma --vectorize-strided-dma ' -O2 --internal-num-neuroncores-per-sengine=1 --logfile=/tmp/nxd_model/token_generation_model/_tp0_bk0/log-neuron-cc.txt --enable-internal-neff-wrapper"
neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:452e29edbf1ade7b85f2c043d9eb590c99124f80447b32b0f8738eda39c1bc36
3
+ size 1604434
neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3864fe0a4d9f6729af3737a4449910c4c4b998013f04cc84758db8054ecc718
3
+ size 7250944
neuronxcc-2.17.194.0+d312836f/MODULE_659e90ae99de96a4c99f+165e9558/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5e7a26e7bc5b24444359b5b3b4d64734f84428428bf5df53a1522fe51ad8d327
3
+ size 7544918
neuronxcc-2.17.194.0+d312836f/MODULE_96d97f202e238feb2c21+bfe5714b/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ "--auto-cast=none --model-type=transformer --tensorizer-options='--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-dge-dma --vectorize-strided-dma ' -O2 --internal-num-neuroncores-per-sengine=1 --logfile=/tmp/nxd_model/context_encoding_model/_tp0_bk0/log-neuron-cc.txt"
neuronxcc-2.17.194.0+d312836f/MODULE_96d97f202e238feb2c21+bfe5714b/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_96d97f202e238feb2c21+bfe5714b/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:160655bcb0e6ff728f6eb2d94d8edd097f723a469cdcac98ef6eef89b96cd0e4
3
+ size 1761757
neuronxcc-2.17.194.0+d312836f/MODULE_96d97f202e238feb2c21+bfe5714b/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e067ecb7dcda6d9fa3a8fd4df832ddc8ab27366a16e835c1c61b7f55c8f82be5
3
+ size 5008384
neuronxcc-2.17.194.0+d312836f/MODULE_9bce7228204bd2d4577b+431f5505/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ "--model-type=transformer -O1 --lnc=1 --internal-hlo2tensorizer-options=--experimental-unsafe-fp8e4m3fn-as-fp8e4m3 --logfile=/tmp/nxd_model/layout_opt/log-neuron-cc.txt"
neuronxcc-2.17.194.0+d312836f/MODULE_9bce7228204bd2d4577b+431f5505/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_9bce7228204bd2d4577b+431f5505/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef4350dbb3041c33ed2ed3737574b565db7f7520f6fb9e34ae56e777eb6b606e
3
+ size 132612
neuronxcc-2.17.194.0+d312836f/MODULE_9bce7228204bd2d4577b+431f5505/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e1cd33146350f57bbd985872991482c18b8c1e2950ec71a402f82baa0f420255
3
+ size 492544
neuronxcc-2.17.194.0+d312836f/MODULE_9cd10f645f9c9838f49a+bfe5714b/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ "--auto-cast=none --model-type=transformer --tensorizer-options='--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-dge-dma --vectorize-strided-dma ' -O2 --internal-num-neuroncores-per-sengine=1 --logfile=/tmp/nxd_model/context_encoding_model/_tp0_bk0/log-neuron-cc.txt"
neuronxcc-2.17.194.0+d312836f/MODULE_9cd10f645f9c9838f49a+bfe5714b/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_9cd10f645f9c9838f49a+bfe5714b/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:118179aa46cbb4bf01184e26d3454300ef0a06de68fe47fcbc22e602c803cec4
3
+ size 823640
neuronxcc-2.17.194.0+d312836f/MODULE_9cd10f645f9c9838f49a+bfe5714b/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c78a3659665e3e35cfc40914a65591bb915ba081fb562bb4116b0300f58ed1c0
3
+ size 27833344