Hugging Face's logo

aws-neuron
/
optimum-neuron-cache

Model card Files
xet
Community
611
dacorvo HF Staff commited on
Commit
c420023
·
verified ·
1 Parent(s): 6ebeb98

Synchronizing local compiler cache.

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +40 -0
  2. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/d337db1f2945982d0bc9.json +58 -0
  3. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/llama/llamafactory/tiny-random-Llama-3/ada00e76710d2d884cc8.json +62 -0
  4. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/llama4_text/tiny-random/llama-4/caf8458124c0c1b8b608.json +81 -0
  5. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/mixtral/dacorvo/Mixtral-tiny/9c1a264c4d265060e661.json +58 -0
  6. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/phi3/yujiepan/phi-4-tiny-random/af5f9033652bae57657c.json +59 -0
  7. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen2/yujiepan/qwen2.5-128k-tiny-random/50dd484535c9c8562031.json +64 -0
  8. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/040b1e23663eba2981b3.json +87 -0
  9. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/6f5f2cf26c21b525ca82.json +87 -0
  10. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/8f2f6fc022fb92a08835.json +87 -0
  11. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/922b8f110b9e3fdaa766.json +87 -0
  12. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-8B/3c013915c647bb8e6712.json +95 -0
  13. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-8B/b60a4e37d9a2dbabe961.json +95 -0
  14. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/39b749f56f6251d4327e.json +65 -0
  15. neuronxcc-2.21.18209.0+043b1bf7/MODULE_06cc3d220bd81f31483c+fb4cc044/compile_flags.json +1 -0
  16. neuronxcc-2.21.18209.0+043b1bf7/MODULE_06cc3d220bd81f31483c+fb4cc044/model.done +0 -0
  17. neuronxcc-2.21.18209.0+043b1bf7/MODULE_06cc3d220bd81f31483c+fb4cc044/model.hlo_module.pb +3 -0
  18. neuronxcc-2.21.18209.0+043b1bf7/MODULE_06cc3d220bd81f31483c+fb4cc044/model.neff +3 -0
  19. neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e5b82b136a38372d842+6e4949b4/compile_flags.json +1 -0
  20. neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e5b82b136a38372d842+6e4949b4/model.done +0 -0
  21. neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e5b82b136a38372d842+6e4949b4/model.hlo_module.pb +3 -0
  22. neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e5b82b136a38372d842+6e4949b4/model.neff +3 -0
  23. neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/compile_flags.json +1 -0
  24. neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/model.done +0 -0
  25. neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/model.hlo_module.pb +3 -0
  26. neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/model.neff +3 -0
  27. neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/wrapped_neff.hlo +3 -0
  28. neuronxcc-2.21.18209.0+043b1bf7/MODULE_315d9491c827c077575a+fb4cc044/compile_flags.json +1 -0
  29. neuronxcc-2.21.18209.0+043b1bf7/MODULE_315d9491c827c077575a+fb4cc044/model.hlo_module.pb +3 -0
  30. neuronxcc-2.21.18209.0+043b1bf7/MODULE_315d9491c827c077575a+fb4cc044/model.log +3 -0
  31. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3321b4e97123c041192d+fb4cc044/compile_flags.json +1 -0
  32. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3321b4e97123c041192d+fb4cc044/model.done +0 -0
  33. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3321b4e97123c041192d+fb4cc044/model.hlo_module.pb +3 -0
  34. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3321b4e97123c041192d+fb4cc044/model.neff +3 -0
  35. neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/compile_flags.json +1 -0
  36. neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/model.done +0 -0
  37. neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/model.hlo_module.pb +3 -0
  38. neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/model.neff +3 -0
  39. neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/wrapped_neff.hlo +3 -0
  40. neuronxcc-2.21.18209.0+043b1bf7/MODULE_394461f5b793e1c6ddbd+fb4cc044/compile_flags.json +1 -0
  41. neuronxcc-2.21.18209.0+043b1bf7/MODULE_394461f5b793e1c6ddbd+fb4cc044/model.done +0 -0
  42. neuronxcc-2.21.18209.0+043b1bf7/MODULE_394461f5b793e1c6ddbd+fb4cc044/model.hlo_module.pb +3 -0
  43. neuronxcc-2.21.18209.0+043b1bf7/MODULE_394461f5b793e1c6ddbd+fb4cc044/model.neff +3 -0
  44. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/compile_flags.json +1 -0
  45. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/model.done +0 -0
  46. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/model.hlo_module.pb +3 -0
  47. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/model.neff +3 -0
  48. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/wrapped_neff.hlo +3 -0
  49. neuronxcc-2.21.18209.0+043b1bf7/MODULE_45c1a0c25eab451751df+fb4cc044/compile_flags.json +1 -0
  50. neuronxcc-2.21.18209.0+043b1bf7/MODULE_45c1a0c25eab451751df+fb4cc044/model.done +0 -0
.gitattributes CHANGED
@@ -12804,3 +12804,43 @@ neuronxcc-2.21.18209.0+043b1bf7/MODULE_1280d8f2f3240abd5752+a02c3a36/wrapped_nef
12804
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_0ca0b7d2c1246aa065a0+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12805
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_d9eacd4b781267f13a69+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12806
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_d9eacd4b781267f13a69+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
12804
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_0ca0b7d2c1246aa065a0+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12805
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_d9eacd4b781267f13a69+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12806
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_d9eacd4b781267f13a69+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12807
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_06cc3d220bd81f31483c+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12808
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e5b82b136a38372d842+6e4949b4/model.neff filter=lfs diff=lfs merge=lfs -text
12809
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12810
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12811
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_3321b4e97123c041192d+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12812
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12813
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12814
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_394461f5b793e1c6ddbd+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12815
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12816
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12817
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_45c1a0c25eab451751df+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12818
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_5988b4b9e2fe27c7105c+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12819
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_5b1c456704a89716d745+2dde74c7/model.neff filter=lfs diff=lfs merge=lfs -text
12820
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_5b1c456704a89716d745+2dde74c7/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12821
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_5f6f4f62aba29e11d946+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12822
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_616029436a1201c3f014+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12823
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_6b37881de0fe95bd8128+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12824
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_823cb759235f53a8fe88+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12825
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_9e174a0ad6384712cfa5+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12826
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_a0387fe4d05bf37e345a+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12827
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_a0387fe4d05bf37e345a+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12828
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_af5ca2287a6b0d70aac7+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12829
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_b559f9b8396895d936b0+80d05c3f/model.neff filter=lfs diff=lfs merge=lfs -text
12830
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_b559f9b8396895d936b0+80d05c3f/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12831
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_bd3408e2c963ac7da056+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12832
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_c345ba838e8c529a43e2+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12833
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_c58e0a47eb485c62b0cd+283df001/model.neff filter=lfs diff=lfs merge=lfs -text
12834
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_cb4d848c61c58d4a5351+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12835
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_ce7eb4c5ef357b61cc09+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12836
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_cedace9d3d693aab8d35+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12837
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_cedace9d3d693aab8d35+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12838
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_d8cfc8fee2dcbc7833f8+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12839
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_da126386d4d4d4a5bf4f+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12840
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_dd4aa1e96fd3b31b162d+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12841
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_de1482f217f05f48d1dd+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
12842
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_e12a32198da9ddff7d98+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12843
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_eaa781288c2e1d7ed527+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12844
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_eaa781288c2e1d7ed527+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12845
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_f4bed0e6c1426c4b9bc7+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12846
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_fd32caf35dc9cdf42fac+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/d337db1f2945982d0bc9.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "GraniteForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "attention_multiplier": 1.0,
11
+ "dtype": "float32",
12
+ "embedding_multiplier": 1.0,
13
+ "hidden_act": "silu",
14
+ "hidden_size": 32,
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 64,
17
+ "logits_scaling": 1.0,
18
+ "max_position_embeddings": 2048,
19
+ "mlp_bias": false,
20
+ "model_type": "granite",
21
+ "neuron": {
22
+ "_serialized_key": "NxDNeuronConfig",
23
+ "batch_size": 1,
24
+ "capacity_factor": null,
25
+ "checkpoint_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
26
+ "checkpoint_revision": "c3074ebc0ac2fe545305f5e5f6cce2cc9b2aa0c5",
27
+ "continuous_batching": false,
28
+ "ep_degree": 1,
29
+ "fused_qkv": true,
30
+ "glu_mlp": true,
31
+ "local_ranks_size": 2,
32
+ "max_batch_size": 1,
33
+ "max_context_length": 1024,
34
+ "max_topk": 256,
35
+ "n_active_tokens": 1024,
36
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.2.dev0",
39
+ "output_logits": false,
40
+ "pp_degree": 1,
41
+ "sequence_length": 1024,
42
+ "speculation_length": 0,
43
+ "start_rank_id": 0,
44
+ "target": "trn1",
45
+ "torch_dtype": "float32",
46
+ "tp_degree": 2
47
+ },
48
+ "num_attention_heads": 4,
49
+ "num_hidden_layers": 2,
50
+ "num_key_value_heads": 4,
51
+ "residual_multiplier": 1.0,
52
+ "rms_norm_eps": 1e-06,
53
+ "rope_scaling": null,
54
+ "rope_theta": 10000.0,
55
+ "tie_word_embeddings": false,
56
+ "use_cache": true,
57
+ "vocab_size": 49152
58
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/llama/llamafactory/tiny-random-Llama-3/ada00e76710d2d884cc8.json ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "llamafactory/tiny-random-Llama-3",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "LlamaForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "float16",
11
+ "head_dim": 4,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 16,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 64,
16
+ "max_position_embeddings": 131072,
17
+ "mlp_bias": false,
18
+ "model_type": "llama",
19
+ "neuron": {
20
+ "_serialized_key": "NxDNeuronConfig",
21
+ "batch_size": 1,
22
+ "capacity_factor": null,
23
+ "checkpoint_id": "llamafactory/tiny-random-Llama-3",
24
+ "checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
25
+ "continuous_batching": false,
26
+ "ep_degree": 1,
27
+ "fused_qkv": true,
28
+ "glu_mlp": true,
29
+ "local_ranks_size": 2,
30
+ "max_batch_size": 1,
31
+ "max_context_length": 1024,
32
+ "max_topk": 256,
33
+ "n_active_tokens": 1024,
34
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.2.dev0",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "sequence_length": 1024,
40
+ "speculation_length": 0,
41
+ "start_rank_id": 0,
42
+ "target": "trn1",
43
+ "torch_dtype": "float16",
44
+ "tp_degree": 2
45
+ },
46
+ "num_attention_heads": 4,
47
+ "num_hidden_layers": 2,
48
+ "num_key_value_heads": 4,
49
+ "pretraining_tp": 1,
50
+ "rms_norm_eps": 1e-05,
51
+ "rope_scaling": {
52
+ "factor": 8.0,
53
+ "high_freq_factor": 4.0,
54
+ "low_freq_factor": 1.0,
55
+ "original_max_position_embeddings": 8192,
56
+ "rope_type": "llama3"
57
+ },
58
+ "rope_theta": 500000.0,
59
+ "tie_word_embeddings": false,
60
+ "use_cache": true,
61
+ "vocab_size": 128256
62
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/llama4_text/tiny-random/llama-4/caf8458124c0c1b8b608.json ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "tiny-random/llama-4",
4
+ "_task": "text-generation",
5
+ "attention_bias": false,
6
+ "attention_chunk_size": 128,
7
+ "attention_dropout": 0.0,
8
+ "attn_scale": 0.1,
9
+ "attn_temperature_tuning": 4,
10
+ "cache_implementation": "hybrid",
11
+ "dtype": "bfloat16",
12
+ "floor_scale": 8192,
13
+ "for_llm_compressor": false,
14
+ "head_dim": 32,
15
+ "hidden_act": "silu",
16
+ "hidden_size": 32,
17
+ "initializer_range": 0.02,
18
+ "interleave_moe_layer_step": 2,
19
+ "intermediate_size": 64,
20
+ "intermediate_size_mlp": 128,
21
+ "layer_types": [
22
+ "chunked_attention",
23
+ "chunked_attention",
24
+ "chunked_attention",
25
+ "full_attention"
26
+ ],
27
+ "max_position_embeddings": 1048576,
28
+ "model_type": "llama4_text",
29
+ "moe_layers": [
30
+ 1,
31
+ 3
32
+ ],
33
+ "neuron": {
34
+ "_serialized_key": "NxDNeuronConfig",
35
+ "batch_size": 1,
36
+ "capacity_factor": null,
37
+ "checkpoint_id": "tiny-random/llama-4",
38
+ "checkpoint_revision": "9e716f5d4d1ffe0a44a15f46f4a12b840439aba4",
39
+ "continuous_batching": false,
40
+ "ep_degree": 1,
41
+ "fused_qkv": false,
42
+ "glu_mlp": true,
43
+ "local_ranks_size": 2,
44
+ "max_batch_size": 1,
45
+ "max_context_length": 1024,
46
+ "max_topk": 256,
47
+ "n_active_tokens": 1024,
48
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
49
+ "on_device_sampling": true,
50
+ "optimum_neuron_version": "0.4.2.dev0",
51
+ "output_logits": false,
52
+ "pp_degree": 1,
53
+ "sequence_length": 1024,
54
+ "speculation_length": 0,
55
+ "start_rank_id": 0,
56
+ "target": "trn1",
57
+ "torch_dtype": "bfloat16",
58
+ "tp_degree": 2
59
+ },
60
+ "no_rope_layers": [
61
+ 1,
62
+ 1,
63
+ 1,
64
+ 0
65
+ ],
66
+ "num_attention_heads": 1,
67
+ "num_experts_per_tok": 1,
68
+ "num_hidden_layers": 4,
69
+ "num_key_value_heads": 1,
70
+ "num_local_experts": 8,
71
+ "output_router_logits": false,
72
+ "rms_norm_eps": 1e-05,
73
+ "rope_scaling": null,
74
+ "rope_theta": 500000.0,
75
+ "router_aux_loss_coef": 0.001,
76
+ "router_jitter_noise": 0.0,
77
+ "tie_word_embeddings": true,
78
+ "use_cache": true,
79
+ "use_qk_norm": true,
80
+ "vocab_size": 202048
81
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/mixtral/dacorvo/Mixtral-tiny/9c1a264c4d265060e661.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "dacorvo/Mixtral-tiny",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "MixtralForCausalLM"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "dtype": "float16",
10
+ "head_dim": 32,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 1024,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 3584,
15
+ "max_position_embeddings": 1024,
16
+ "model_type": "mixtral",
17
+ "neuron": {
18
+ "_serialized_key": "NxDNeuronConfig",
19
+ "batch_size": 1,
20
+ "capacity_factor": null,
21
+ "checkpoint_id": "dacorvo/Mixtral-tiny",
22
+ "checkpoint_revision": "c557ba205ddff6ea911f4719e0d543d6c08356b6",
23
+ "continuous_batching": false,
24
+ "ep_degree": 1,
25
+ "fused_qkv": false,
26
+ "glu_mlp": true,
27
+ "local_ranks_size": 2,
28
+ "max_batch_size": 1,
29
+ "max_context_length": 1024,
30
+ "max_topk": 256,
31
+ "n_active_tokens": 1024,
32
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
33
+ "on_device_sampling": false,
34
+ "optimum_neuron_version": "0.4.2.dev0",
35
+ "output_logits": false,
36
+ "pp_degree": 1,
37
+ "sequence_length": 1024,
38
+ "speculation_length": 0,
39
+ "start_rank_id": 0,
40
+ "target": "trn1",
41
+ "torch_dtype": "float16",
42
+ "tp_degree": 2
43
+ },
44
+ "num_attention_heads": 32,
45
+ "num_experts_per_tok": 2,
46
+ "num_hidden_layers": 2,
47
+ "num_key_value_heads": 8,
48
+ "num_local_experts": 8,
49
+ "output_router_logits": false,
50
+ "rms_norm_eps": 1e-05,
51
+ "rope_theta": 10000.0,
52
+ "router_aux_loss_coef": 0.001,
53
+ "router_jitter_noise": 0.0,
54
+ "sliding_window": 4096,
55
+ "tie_word_embeddings": false,
56
+ "use_cache": true,
57
+ "vocab_size": 32000
58
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/phi3/yujiepan/phi-4-tiny-random/af5f9033652bae57657c.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "yujiepan/phi-4-tiny-random",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Phi3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "auto_map": {},
11
+ "dtype": "bfloat16",
12
+ "embd_pdrop": 0.0,
13
+ "hidden_act": "silu",
14
+ "hidden_size": 16,
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 32,
17
+ "max_position_embeddings": 16384,
18
+ "model_type": "phi3",
19
+ "neuron": {
20
+ "_serialized_key": "NxDNeuronConfig",
21
+ "batch_size": 1,
22
+ "capacity_factor": null,
23
+ "checkpoint_id": "yujiepan/phi-4-tiny-random",
24
+ "checkpoint_revision": "18a9a1168dc97ac6d128f811925670c275610f5a",
25
+ "continuous_batching": false,
26
+ "ep_degree": 1,
27
+ "fused_qkv": true,
28
+ "glu_mlp": true,
29
+ "local_ranks_size": 2,
30
+ "max_batch_size": 1,
31
+ "max_context_length": 1024,
32
+ "max_topk": 256,
33
+ "n_active_tokens": 1024,
34
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.2.dev0",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "sequence_length": 1024,
40
+ "speculation_length": 0,
41
+ "start_rank_id": 0,
42
+ "target": "trn1",
43
+ "torch_dtype": "bfloat16",
44
+ "tp_degree": 2
45
+ },
46
+ "num_attention_heads": 2,
47
+ "num_hidden_layers": 2,
48
+ "num_key_value_heads": 1,
49
+ "original_max_position_embeddings": 16384,
50
+ "partial_rotary_factor": 1.0,
51
+ "resid_pdrop": 0.0,
52
+ "rms_norm_eps": 1e-05,
53
+ "rope_scaling": null,
54
+ "rope_theta": 250000,
55
+ "sliding_window": null,
56
+ "tie_word_embeddings": false,
57
+ "use_cache": true,
58
+ "vocab_size": 100352
59
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen2/yujiepan/qwen2.5-128k-tiny-random/50dd484535c9c8562031.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "yujiepan/qwen2.5-128k-tiny-random",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen2ForCausalLM"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "dtype": "bfloat16",
10
+ "hidden_act": "silu",
11
+ "hidden_size": 8,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 16,
14
+ "layer_types": [
15
+ "full_attention",
16
+ "full_attention"
17
+ ],
18
+ "max_position_embeddings": 32768,
19
+ "max_window_layers": 1,
20
+ "model_type": "qwen2",
21
+ "neuron": {
22
+ "_serialized_key": "NxDNeuronConfig",
23
+ "batch_size": 1,
24
+ "capacity_factor": null,
25
+ "checkpoint_id": "yujiepan/qwen2.5-128k-tiny-random",
26
+ "checkpoint_revision": "c8296d4ca3f87782876d2382fbb6481d1beb8ef0",
27
+ "continuous_batching": false,
28
+ "ep_degree": 1,
29
+ "fused_qkv": false,
30
+ "glu_mlp": true,
31
+ "local_ranks_size": 2,
32
+ "max_batch_size": 1,
33
+ "max_context_length": 1024,
34
+ "max_topk": 256,
35
+ "n_active_tokens": 1024,
36
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.2.dev0",
39
+ "output_logits": false,
40
+ "pp_degree": 1,
41
+ "sequence_length": 1024,
42
+ "speculation_length": 0,
43
+ "start_rank_id": 0,
44
+ "target": "trn1",
45
+ "torch_dtype": "bfloat16",
46
+ "tp_degree": 2
47
+ },
48
+ "num_attention_heads": 4,
49
+ "num_hidden_layers": 2,
50
+ "num_key_value_heads": 2,
51
+ "rms_norm_eps": 1e-06,
52
+ "rope_scaling": {
53
+ "factor": 4.0,
54
+ "original_max_position_embeddings": 32768,
55
+ "rope_type": "yarn",
56
+ "type": "yarn"
57
+ },
58
+ "rope_theta": 1000000.0,
59
+ "sliding_window": null,
60
+ "tie_word_embeddings": false,
61
+ "use_cache": true,
62
+ "use_sliding_window": false,
63
+ "vocab_size": 152064
64
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/040b1e23663eba2981b3.json ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen3-Embedding-0.6B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 128,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 1024,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 3072,
16
+ "layer_types": [
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention"
45
+ ],
46
+ "max_position_embeddings": 32768,
47
+ "max_window_layers": 28,
48
+ "model_type": "qwen3",
49
+ "neuron": {
50
+ "_serialized_key": "NxDNeuronConfig",
51
+ "batch_size": 6,
52
+ "capacity_factor": null,
53
+ "checkpoint_id": "Qwen/Qwen3-Embedding-0.6B",
54
+ "checkpoint_revision": "c54f2e6e80b2d7b7de06f51cec4959f6b3e03418",
55
+ "continuous_batching": false,
56
+ "ep_degree": 1,
57
+ "fused_qkv": true,
58
+ "glu_mlp": true,
59
+ "local_ranks_size": 24,
60
+ "max_batch_size": 6,
61
+ "max_context_length": 1024,
62
+ "max_topk": 256,
63
+ "n_active_tokens": 1024,
64
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
65
+ "on_device_sampling": false,
66
+ "optimum_neuron_version": "0.4.2.dev0",
67
+ "output_logits": false,
68
+ "pp_degree": 1,
69
+ "sequence_length": 1024,
70
+ "speculation_length": 0,
71
+ "start_rank_id": 0,
72
+ "target": "trn1",
73
+ "torch_dtype": "bfloat16",
74
+ "tp_degree": 24
75
+ },
76
+ "num_attention_heads": 16,
77
+ "num_hidden_layers": 28,
78
+ "num_key_value_heads": 8,
79
+ "rms_norm_eps": 1e-06,
80
+ "rope_scaling": null,
81
+ "rope_theta": 1000000,
82
+ "sliding_window": null,
83
+ "tie_word_embeddings": true,
84
+ "use_cache": true,
85
+ "use_sliding_window": false,
86
+ "vocab_size": 151669
87
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/6f5f2cf26c21b525ca82.json ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen3-Embedding-0.6B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 128,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 1024,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 3072,
16
+ "layer_types": [
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention"
45
+ ],
46
+ "max_position_embeddings": 32768,
47
+ "max_window_layers": 28,
48
+ "model_type": "qwen3",
49
+ "neuron": {
50
+ "_serialized_key": "NxDNeuronConfig",
51
+ "batch_size": 1,
52
+ "capacity_factor": null,
53
+ "checkpoint_id": "Qwen/Qwen3-Embedding-0.6B",
54
+ "checkpoint_revision": "c54f2e6e80b2d7b7de06f51cec4959f6b3e03418",
55
+ "continuous_batching": false,
56
+ "ep_degree": 1,
57
+ "fused_qkv": true,
58
+ "glu_mlp": true,
59
+ "local_ranks_size": 1,
60
+ "max_batch_size": 1,
61
+ "max_context_length": 4096,
62
+ "max_topk": 256,
63
+ "n_active_tokens": 4096,
64
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
65
+ "on_device_sampling": false,
66
+ "optimum_neuron_version": "0.4.2.dev0",
67
+ "output_logits": false,
68
+ "pp_degree": 1,
69
+ "sequence_length": 4096,
70
+ "speculation_length": 0,
71
+ "start_rank_id": 0,
72
+ "target": "trn1",
73
+ "torch_dtype": "bfloat16",
74
+ "tp_degree": 1
75
+ },
76
+ "num_attention_heads": 16,
77
+ "num_hidden_layers": 28,
78
+ "num_key_value_heads": 8,
79
+ "rms_norm_eps": 1e-06,
80
+ "rope_scaling": null,
81
+ "rope_theta": 1000000,
82
+ "sliding_window": null,
83
+ "tie_word_embeddings": true,
84
+ "use_cache": true,
85
+ "use_sliding_window": false,
86
+ "vocab_size": 151669
87
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/8f2f6fc022fb92a08835.json ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen3-Embedding-0.6B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 128,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 1024,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 3072,
16
+ "layer_types": [
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention"
45
+ ],
46
+ "max_position_embeddings": 32768,
47
+ "max_window_layers": 28,
48
+ "model_type": "qwen3",
49
+ "neuron": {
50
+ "_serialized_key": "NxDNeuronConfig",
51
+ "batch_size": 4,
52
+ "capacity_factor": null,
53
+ "checkpoint_id": "Qwen/Qwen3-Embedding-0.6B",
54
+ "checkpoint_revision": "c54f2e6e80b2d7b7de06f51cec4959f6b3e03418",
55
+ "continuous_batching": false,
56
+ "ep_degree": 1,
57
+ "fused_qkv": true,
58
+ "glu_mlp": true,
59
+ "local_ranks_size": 24,
60
+ "max_batch_size": 4,
61
+ "max_context_length": 1024,
62
+ "max_topk": 256,
63
+ "n_active_tokens": 1024,
64
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
65
+ "on_device_sampling": false,
66
+ "optimum_neuron_version": "0.4.2.dev0",
67
+ "output_logits": false,
68
+ "pp_degree": 1,
69
+ "sequence_length": 1024,
70
+ "speculation_length": 0,
71
+ "start_rank_id": 0,
72
+ "target": "trn1",
73
+ "torch_dtype": "bfloat16",
74
+ "tp_degree": 24
75
+ },
76
+ "num_attention_heads": 16,
77
+ "num_hidden_layers": 28,
78
+ "num_key_value_heads": 8,
79
+ "rms_norm_eps": 1e-06,
80
+ "rope_scaling": null,
81
+ "rope_theta": 1000000,
82
+ "sliding_window": null,
83
+ "tie_word_embeddings": true,
84
+ "use_cache": true,
85
+ "use_sliding_window": false,
86
+ "vocab_size": 151669
87
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/922b8f110b9e3fdaa766.json ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen3-Embedding-0.6B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 128,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 1024,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 3072,
16
+ "layer_types": [
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention"
45
+ ],
46
+ "max_position_embeddings": 32768,
47
+ "max_window_layers": 28,
48
+ "model_type": "qwen3",
49
+ "neuron": {
50
+ "_serialized_key": "NxDNeuronConfig",
51
+ "batch_size": 1,
52
+ "capacity_factor": null,
53
+ "checkpoint_id": "Qwen/Qwen3-Embedding-0.6B",
54
+ "checkpoint_revision": "c54f2e6e80b2d7b7de06f51cec4959f6b3e03418",
55
+ "continuous_batching": false,
56
+ "ep_degree": 1,
57
+ "fused_qkv": true,
58
+ "glu_mlp": true,
59
+ "local_ranks_size": 24,
60
+ "max_batch_size": 1,
61
+ "max_context_length": 1024,
62
+ "max_topk": 256,
63
+ "n_active_tokens": 1024,
64
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
65
+ "on_device_sampling": false,
66
+ "optimum_neuron_version": "0.4.2.dev0",
67
+ "output_logits": false,
68
+ "pp_degree": 1,
69
+ "sequence_length": 1024,
70
+ "speculation_length": 0,
71
+ "start_rank_id": 0,
72
+ "target": "trn1",
73
+ "torch_dtype": "bfloat16",
74
+ "tp_degree": 24
75
+ },
76
+ "num_attention_heads": 16,
77
+ "num_hidden_layers": 28,
78
+ "num_key_value_heads": 8,
79
+ "rms_norm_eps": 1e-06,
80
+ "rope_scaling": null,
81
+ "rope_theta": 1000000,
82
+ "sliding_window": null,
83
+ "tie_word_embeddings": true,
84
+ "use_cache": true,
85
+ "use_sliding_window": false,
86
+ "vocab_size": 151669
87
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-8B/3c013915c647bb8e6712.json ADDED
@@ -0,0 +1,95 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen3-Embedding-8B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 128,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 4096,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 12288,
16
+ "layer_types": [
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention",
45
+ "full_attention",
46
+ "full_attention",
47
+ "full_attention",
48
+ "full_attention",
49
+ "full_attention",
50
+ "full_attention",
51
+ "full_attention",
52
+ "full_attention"
53
+ ],
54
+ "max_position_embeddings": 40960,
55
+ "max_window_layers": 36,
56
+ "model_type": "qwen3",
57
+ "neuron": {
58
+ "_serialized_key": "NxDNeuronConfig",
59
+ "batch_size": 1,
60
+ "capacity_factor": null,
61
+ "checkpoint_id": "Qwen/Qwen3-Embedding-8B",
62
+ "checkpoint_revision": "1d8ad4ca9b3dd8059ad90a75d4983776a23d44af",
63
+ "continuous_batching": false,
64
+ "ep_degree": 1,
65
+ "fused_qkv": true,
66
+ "glu_mlp": true,
67
+ "local_ranks_size": 2,
68
+ "max_batch_size": 1,
69
+ "max_context_length": 4096,
70
+ "max_topk": 256,
71
+ "n_active_tokens": 4096,
72
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
73
+ "on_device_sampling": false,
74
+ "optimum_neuron_version": "0.4.2.dev0",
75
+ "output_logits": false,
76
+ "pp_degree": 1,
77
+ "sequence_length": 4096,
78
+ "speculation_length": 0,
79
+ "start_rank_id": 0,
80
+ "target": "trn1",
81
+ "torch_dtype": "bfloat16",
82
+ "tp_degree": 2
83
+ },
84
+ "num_attention_heads": 32,
85
+ "num_hidden_layers": 36,
86
+ "num_key_value_heads": 8,
87
+ "rms_norm_eps": 1e-06,
88
+ "rope_scaling": null,
89
+ "rope_theta": 1000000,
90
+ "sliding_window": null,
91
+ "tie_word_embeddings": false,
92
+ "use_cache": true,
93
+ "use_sliding_window": false,
94
+ "vocab_size": 151665
95
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3/Qwen/Qwen3-Embedding-8B/b60a4e37d9a2dbabe961.json ADDED
@@ -0,0 +1,95 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen3-Embedding-8B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 128,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 4096,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 12288,
16
+ "layer_types": [
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention",
45
+ "full_attention",
46
+ "full_attention",
47
+ "full_attention",
48
+ "full_attention",
49
+ "full_attention",
50
+ "full_attention",
51
+ "full_attention",
52
+ "full_attention"
53
+ ],
54
+ "max_position_embeddings": 40960,
55
+ "max_window_layers": 36,
56
+ "model_type": "qwen3",
57
+ "neuron": {
58
+ "_serialized_key": "NxDNeuronConfig",
59
+ "batch_size": 16,
60
+ "capacity_factor": null,
61
+ "checkpoint_id": "Qwen/Qwen3-Embedding-8B",
62
+ "checkpoint_revision": "1d8ad4ca9b3dd8059ad90a75d4983776a23d44af",
63
+ "continuous_batching": false,
64
+ "ep_degree": 1,
65
+ "fused_qkv": true,
66
+ "glu_mlp": true,
67
+ "local_ranks_size": 2,
68
+ "max_batch_size": 16,
69
+ "max_context_length": 4096,
70
+ "max_topk": 256,
71
+ "n_active_tokens": 4096,
72
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
73
+ "on_device_sampling": false,
74
+ "optimum_neuron_version": "0.4.2.dev0",
75
+ "output_logits": false,
76
+ "pp_degree": 1,
77
+ "sequence_length": 4096,
78
+ "speculation_length": 0,
79
+ "start_rank_id": 0,
80
+ "target": "trn1",
81
+ "torch_dtype": "bfloat16",
82
+ "tp_degree": 2
83
+ },
84
+ "num_attention_heads": 32,
85
+ "num_hidden_layers": 36,
86
+ "num_key_value_heads": 8,
87
+ "rms_norm_eps": 1e-06,
88
+ "rope_scaling": null,
89
+ "rope_theta": 1000000,
90
+ "sliding_window": null,
91
+ "tie_word_embeddings": false,
92
+ "use_cache": true,
93
+ "use_sliding_window": false,
94
+ "vocab_size": 151665
95
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev0/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/39b749f56f6251d4327e.json ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "optimum-internal-testing/tiny-random-qwen3_moe",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3MoeForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "decoder_sparse_step": 2,
11
+ "dtype": "float32",
12
+ "head_dim": 32,
13
+ "hidden_act": "silu",
14
+ "hidden_size": 64,
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 128,
17
+ "max_position_embeddings": 40960,
18
+ "max_window_layers": 1,
19
+ "mlp_only_layers": [],
20
+ "model_type": "qwen3_moe",
21
+ "moe_intermediate_size": 128,
22
+ "neuron": {
23
+ "_serialized_key": "NxDNeuronConfig",
24
+ "batch_size": 1,
25
+ "capacity_factor": null,
26
+ "checkpoint_id": "optimum-internal-testing/tiny-random-qwen3_moe",
27
+ "checkpoint_revision": "e0230be2839556b44b7400a233c73c74b4abb7af",
28
+ "continuous_batching": false,
29
+ "ep_degree": 1,
30
+ "fused_qkv": false,
31
+ "glu_mlp": true,
32
+ "local_ranks_size": 2,
33
+ "max_batch_size": 1,
34
+ "max_context_length": 1024,
35
+ "max_topk": 256,
36
+ "n_active_tokens": 1024,
37
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
38
+ "on_device_sampling": true,
39
+ "optimum_neuron_version": "0.4.2.dev0",
40
+ "output_logits": false,
41
+ "pp_degree": 1,
42
+ "sequence_length": 1024,
43
+ "speculation_length": 0,
44
+ "start_rank_id": 0,
45
+ "target": "trn1",
46
+ "torch_dtype": "float32",
47
+ "tp_degree": 2
48
+ },
49
+ "norm_topk_prob": true,
50
+ "num_attention_heads": 2,
51
+ "num_experts": 8,
52
+ "num_experts_per_tok": 2,
53
+ "num_hidden_layers": 2,
54
+ "num_key_value_heads": 1,
55
+ "output_router_logits": false,
56
+ "rms_norm_eps": 1e-06,
57
+ "rope_scaling": null,
58
+ "rope_theta": 1000000.0,
59
+ "router_aux_loss_coef": 0.001,
60
+ "sliding_window": null,
61
+ "tie_word_embeddings": true,
62
+ "use_cache": true,
63
+ "use_sliding_window": false,
64
+ "vocab_size": 151936
65
+ }
neuronxcc-2.21.18209.0+043b1bf7/MODULE_06cc3d220bd81f31483c+fb4cc044/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/encoding/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_06cc3d220bd81f31483c+fb4cc044/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_06cc3d220bd81f31483c+fb4cc044/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7607df5f2c7c5132d87730619aa3366ddb2035e223a239aa5a14a1c9d7469d57
3
+ size 848773
neuronxcc-2.21.18209.0+043b1bf7/MODULE_06cc3d220bd81f31483c+fb4cc044/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7d799faaff5e25a7531d12f187cea88faf79b29e045c2b88d2c3ea0b8bd94cab
3
+ size 5786624
neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e5b82b136a38372d842+6e4949b4/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--enable-saturate-infinity", "--enable-mixed-precision-accumulation", "--model-type", "transformer", "-O1", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2", "--auto-cast=none", "--internal-enable-dge-levels", "vector_dynamic_offsets", "--logfile=/tmp/nxd_model/context_encoding/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e5b82b136a38372d842+6e4949b4/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e5b82b136a38372d842+6e4949b4/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ceb0a23b716925f843a16e15f9f0385f067950c8f9f55cd7b3db583ba03b1a7e
3
+ size 84807
neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e5b82b136a38372d842+6e4949b4/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0f26bc2e40e6e21bffa40348dd28785a0c5ba6ae26126553c85435559849af5d
3
+ size 646144
neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/token_generation/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d61ac2165fde7772e8d3035b0f89b50beec5e0d305548cbacc624eefa9452d3
3
+ size 82753
neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:690971e38e78e6e8e2e0b2249904244dbe845f18a1d36cb2b763deac5fe550f3
3
+ size 277504
neuronxcc-2.21.18209.0+043b1bf7/MODULE_306829ff9e3811933fe7+a02c3a36/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:937aba417c7465954a0bf71e8d536e6ce2db41d44b477deb724f8f6154f76e19
3
+ size 285854
neuronxcc-2.21.18209.0+043b1bf7/MODULE_315d9491c827c077575a+fb4cc044/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/encoding/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_315d9491c827c077575a+fb4cc044/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:949343b64aaeb75ac893058b4181a58759ab30c14ab879041244bb9af4356695
3
+ size 850786
neuronxcc-2.21.18209.0+043b1bf7/MODULE_315d9491c827c077575a+fb4cc044/model.log ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ Failed compilation with ['neuronx-cc', 'compile', '--framework=XLA', '/tmp/nxd_model/encoding/_tp0_bk0/model.MODULE_315d9491c827c077575a+fb4cc044.hlo_module.pb', '--output', '/tmp/nxd_model/encoding/_tp0_bk0/model.MODULE_315d9491c827c077575a+fb4cc044.neff', '--target=trn1', '--auto-cast=none', '--model-type=transformer', '--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ', '-O2', '--lnc=1', '--logfile=/tmp/nxd_model/encoding/_tp0_bk0/log-neuron-cc.txt', '--verbose=35']: [XCG815] Estimated peak HBM usage (19.257GB) exceeds 16GB. Neff might be unable to load on chip. If you believe this estimation to be inaccurate, you can disable the check using: `--internal-backend-options=' --disable-hbm-usage-check '` - Please open a support ticket at https://github.com/aws-neuron/aws-neuron-sdk/issues/new. You may also be able to obtain more information using the 'XLA_IR_DEBUG' and 'XLA_HLO_DEBUG' environment variables.
2
+ 2025-11-07T14:10:10Z Non-signal exit. Backend exited with code 1 and stderr: [XCG815] Estimated peak HBM usage (19.257GB) exceeds 16GB. Neff might be unable to load on chip. If you believe this estimation to be inaccurate, you can disable the check using: `--internal-backend-options=' --disable-hbm-usage-check '` - Please open a support ticket at https://github.com/aws-neuron/aws-neuron-sdk/issues/new. You may also be able to obtain more information using the 'XLA_IR_DEBUG' and 'XLA_HLO_DEBUG' environment variables.
3
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3321b4e97123c041192d+fb4cc044/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/encoding/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3321b4e97123c041192d+fb4cc044/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3321b4e97123c041192d+fb4cc044/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34e6c39ae5aaa2365e0018b111269da40c42b131faed18711a8810bedd68a436
3
+ size 429397
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3321b4e97123c041192d+fb4cc044/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d0135e8fa145738be49848914c2dbb2c52c3854a2403644f1842244e432ed0fe
3
+ size 1352704
neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/token_generation/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:20ed0858162f8fc2c58bc4c23d0925dbeda5897dae0889f87758206676394f4e
3
+ size 81843
neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:adc574992e72adfe26c916de4e344ec745c5f0367620c1183437c3127f9fef15
3
+ size 246784
neuronxcc-2.21.18209.0+043b1bf7/MODULE_34362baee2ad4784108e+a02c3a36/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6e3995a13290761f8f6f89854871f6a5da48002e637efad7b67cec72e826cc1e
3
+ size 255104
neuronxcc-2.21.18209.0+043b1bf7/MODULE_394461f5b793e1c6ddbd+fb4cc044/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/encoding/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_394461f5b793e1c6ddbd+fb4cc044/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_394461f5b793e1c6ddbd+fb4cc044/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a36ee034700342382b6da2932f74a5d63b34f8b66d47da8dcd5b0cf42842dd7
3
+ size 428915
neuronxcc-2.21.18209.0+043b1bf7/MODULE_394461f5b793e1c6ddbd+fb4cc044/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2978f644a5df5278f76096838837e039d1d5a3dc2cf4e9c2a6cae1784f9bec0
3
+ size 4199424
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/token_generation/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e70386c0d4c7acaf312dfe4908f119fe46d817ab3d82394d549c5ed84707d551
3
+ size 80382
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:40684a5208697e733fd92b531e426c9a49e2e2f3dce8fdc81afe434818f265f2
3
+ size 216064
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3e3c9e9db5ebafa6c488+a02c3a36/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e490cb2df34ae3582e58811ed1a9ea678ca780e9525b51113490839860539904
3
+ size 224412
neuronxcc-2.21.18209.0+043b1bf7/MODULE_45c1a0c25eab451751df+fb4cc044/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/encoding/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_45c1a0c25eab451751df+fb4cc044/model.done ADDED
File without changes