dacorvo HF Staff commited on
Commit
700ac84
·
verified ·
1 Parent(s): ab5eea5

Synchronizing local compiler cache.

Browse files
Files changed (41) hide show
  1. .gitattributes +5 -0
  2. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/03b0c107d1cede36875199a5d51decfe04c473de2af9999f8577a028d74d0ab4/2339cde11cbd17accc62.json +189 -0
  3. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/2da7a00f0478d50ae1e7f75f085c5b2773b5f355f427c61cf34cb6febd629d96/7d7f3f7850d2aa89e3fa.json +60 -0
  4. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/441269935591cad8d370e512c0b93cdd2fce6247c40e5a4866d872ee5338b0de/01957724439347680e1b.json +82 -0
  5. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/4ab8140bc7eb4a553d95855c5c2be2cf8c0fbab21b823d76183b6f51e98b6fc5/80038e075c3313d11a1d.json +59 -0
  6. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/4cb7aff9e2a15c151396f2b684013e39d6739f0dec83e5c9dabbfe9d5fcf77b7/e28a26007281cdef7599.json +83 -0
  7. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/6454afdf3e9d66c7226c13a575b718845c25e53b0699600ba2bb4f883e9d841b/b2922eb58f0f3319bfe3.json +63 -0
  8. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/73707b485eab9008c7aba7f5dad0ce2384ac685318d5f888c12fa0d81ed90b19/3aed09186384cc6540ef.json +135 -0
  9. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/7518518c7e077820070186deda960d8cc49db068cdf0ac70664098fa2b6b698c/c7e74a139ce290675aa6.json +65 -0
  10. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/7f05bde17c7b0ffeb657897697f23d182f406b76ced7f1b2cd5741dc93fe2e2e/ca71d7a47e38a69da803.json +126 -0
  11. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/920f44ce6d3e004d1ce547ae06644f7be262180644b04573153aa15d98742edc/d6743412c8cf7830ae4a.json +66 -0
  12. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/cf6b9a360dcf294104671106bae2adbd9fd291823bb60a351883163684073231/7d57cf017fab771c9e95.json +64 -0
  13. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/cf6b9a360dcf294104671106bae2adbd9fd291823bb60a351883163684073231/9fdf2fe50b15951e2b95.json +64 -0
  14. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/cf6b9a360dcf294104671106bae2adbd9fd291823bb60a351883163684073231/c737e0d82a12fca4c663.json +64 -0
  15. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/d139acf64685f15794bb983ff6eb881bdd31304bae88b0ce1ed20a54c21f2265/fcca5755ba87a7768913.json +59 -0
  16. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/gemma3_text/unsloth/gemma-3-270m-it/01957724439347680e1b.json +82 -0
  17. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/fcca5755ba87a7768913.json +59 -0
  18. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/idefics3/HuggingFaceTB/SmolVLM-256M-Instruct/2339cde11cbd17accc62.json +189 -0
  19. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/llama/llamafactory/tiny-random-Llama-3/b2922eb58f0f3319bfe3.json +63 -0
  20. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/llama/unsloth/Llama-3.2-1B-Instruct/c737e0d82a12fca4c663.json +64 -0
  21. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/llama4/tiny-random/llama-4/ca71d7a47e38a69da803.json +126 -0
  22. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/mixtral/dacorvo/Mixtral-tiny/80038e075c3313d11a1d.json +59 -0
  23. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/phi3/yujiepan/phi-4-tiny-random/7d7f3f7850d2aa89e3fa.json +60 -0
  24. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/qwen2/Qwen/Qwen2.5-0.5B/e28a26007281cdef7599.json +83 -0
  25. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/qwen2/yujiepan/qwen2.5-128k-tiny-random/c7e74a139ce290675aa6.json +65 -0
  26. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/d6743412c8cf7830ae4a.json +66 -0
  27. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/smollm3/HuggingFaceTB/SmolLM3-3B/3aed09186384cc6540ef.json +135 -0
  28. neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/compile_flags.json +1 -0
  29. neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/model.done +0 -0
  30. neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/model.hlo_module.pb +3 -0
  31. neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/model.neff +3 -0
  32. neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/wrapped_neff.hlo +3 -0
  33. neuronxcc-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+a02c3a36/compile_flags.json +1 -0
  34. neuronxcc-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+a02c3a36/model.done +0 -0
  35. neuronxcc-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+a02c3a36/model.hlo_module.pb +3 -0
  36. neuronxcc-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+a02c3a36/model.neff +3 -0
  37. neuronxcc-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+a02c3a36/wrapped_neff.hlo +3 -0
  38. neuronxcc-2.21.33363.0+82129205/MODULE_82382ce0b34b3e272481+6170d8e1/compile_flags.json +1 -0
  39. neuronxcc-2.21.33363.0+82129205/MODULE_82382ce0b34b3e272481+6170d8e1/model.done +0 -0
  40. neuronxcc-2.21.33363.0+82129205/MODULE_82382ce0b34b3e272481+6170d8e1/model.hlo_module.pb +3 -0
  41. neuronxcc-2.21.33363.0+82129205/MODULE_82382ce0b34b3e272481+6170d8e1/model.neff +3 -0
.gitattributes CHANGED
@@ -17399,3 +17399,8 @@ neuronxcc-2.21.33363.0+82129205/MODULE_880bcc864ca6414a5589+82ae0560/model.neff
17399
  neuronxcc-2.21.33363.0+82129205/MODULE_880bcc864ca6414a5589+82ae0560/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
17400
  neuronxcc-2.21.33363.0+82129205/MODULE_f850a063c4798248da2d+2a75dc25/model.neff filter=lfs diff=lfs merge=lfs -text
17401
  neuronxcc-2.21.33363.0+82129205/MODULE_f850a063c4798248da2d+2a75dc25/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
17399
  neuronxcc-2.21.33363.0+82129205/MODULE_880bcc864ca6414a5589+82ae0560/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
17400
  neuronxcc-2.21.33363.0+82129205/MODULE_f850a063c4798248da2d+2a75dc25/model.neff filter=lfs diff=lfs merge=lfs -text
17401
  neuronxcc-2.21.33363.0+82129205/MODULE_f850a063c4798248da2d+2a75dc25/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
17402
+ neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/model.neff filter=lfs diff=lfs merge=lfs -text
17403
+ neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
17404
+ neuronxcc-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
17405
+ neuronxcc-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
17406
+ neuronxcc-2.21.33363.0+82129205/MODULE_82382ce0b34b3e272481+6170d8e1/model.neff filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/03b0c107d1cede36875199a5d51decfe04c473de2af9999f8577a028d74d0ab4/2339cde11cbd17accc62.json ADDED
@@ -0,0 +1,189 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "HuggingFaceTB/SmolVLM-256M-Instruct",
4
+ "_task": "image-text-to-text",
5
+ "architectures": [
6
+ "Idefics3ForConditionalGeneration"
7
+ ],
8
+ "dtype": "bfloat16",
9
+ "image_token_id": 49190,
10
+ "model_type": "idefics3",
11
+ "neuron": {
12
+ "_serialized_key": "NxDVLMNeuronConfig",
13
+ "batch_size": 1,
14
+ "capacity_factor": null,
15
+ "checkpoint_id": "HuggingFaceTB/SmolVLM-256M-Instruct",
16
+ "checkpoint_revision": "7e3e67edbbed1bf9888184d9df282b700a323964",
17
+ "continuous_batching": false,
18
+ "ep_degree": 1,
19
+ "fused_qkv": true,
20
+ "glu_mlp": true,
21
+ "image_seq_len": 64,
22
+ "image_size": 512,
23
+ "local_ranks_size": 2,
24
+ "max_batch_size": 1,
25
+ "max_context_length": 2048,
26
+ "max_num_images": 17,
27
+ "max_topk": 256,
28
+ "n_active_tokens": 2048,
29
+ "neuronxcc_version": "2.21.33363.0+82129205",
30
+ "on_device_sampling": true,
31
+ "optimum_neuron_version": "0.4.6.dev4",
32
+ "output_logits": false,
33
+ "pp_degree": 1,
34
+ "prefill_chunk_size": 1024,
35
+ "sequence_length": 2048,
36
+ "speculation_length": 0,
37
+ "start_rank_id": 0,
38
+ "target": "trn1",
39
+ "torch_dtype": "bfloat16",
40
+ "tp_degree": 2
41
+ },
42
+ "scale_factor": 4,
43
+ "text_config": {
44
+ "_attn_implementation_autoset": false,
45
+ "_flash_attn_2_enabled": true,
46
+ "_name_or_path": "None",
47
+ "architectures": [
48
+ "VLlama3ForCausalLM"
49
+ ],
50
+ "attention_bias": false,
51
+ "attention_dropout": 0.0,
52
+ "dtype": "bfloat16",
53
+ "head_dim": 64,
54
+ "hidden_act": "silu",
55
+ "hidden_size": 576,
56
+ "initializer_range": 0.041666666666666664,
57
+ "intermediate_size": 1536,
58
+ "is_llama_config": true,
59
+ "max_position_embeddings": 8192,
60
+ "mlp_bias": false,
61
+ "model_type": "llama",
62
+ "neftune_noise_alpha": 0.0,
63
+ "num_attention_heads": 9,
64
+ "num_hidden_layers": 30,
65
+ "num_key_value_heads": 3,
66
+ "pad_token_id": 2,
67
+ "perceiver_config": {
68
+ "_attn_implementation_autoset": false,
69
+ "_name_or_path": "",
70
+ "add_cross_attention": false,
71
+ "architectures": null,
72
+ "attention_dropout": 0.0,
73
+ "bad_words_ids": null,
74
+ "begin_suppress_tokens": null,
75
+ "bos_token_id": null,
76
+ "chunk_size_feed_forward": 0,
77
+ "cross_attention_hidden_size": null,
78
+ "decoder_start_token_id": null,
79
+ "diversity_penalty": 0.0,
80
+ "do_sample": false,
81
+ "early_stopping": false,
82
+ "encoder_no_repeat_ngram_size": 0,
83
+ "eos_token_id": null,
84
+ "exponential_decay_length_penalty": null,
85
+ "finetuning_task": null,
86
+ "forced_bos_token_id": null,
87
+ "forced_eos_token_id": null,
88
+ "hidden_act": "silu",
89
+ "id2label": {
90
+ "0": "LABEL_0",
91
+ "1": "LABEL_1"
92
+ },
93
+ "is_decoder": false,
94
+ "is_encoder_decoder": false,
95
+ "label2id": {
96
+ "LABEL_0": 0,
97
+ "LABEL_1": 1
98
+ },
99
+ "length_penalty": 1.0,
100
+ "max_length": 20,
101
+ "min_length": 0,
102
+ "model_type": "vllama3",
103
+ "no_repeat_ngram_size": 0,
104
+ "num_beam_groups": 1,
105
+ "num_beams": 1,
106
+ "num_key_value_heads": 1,
107
+ "num_return_sequences": 1,
108
+ "output_attentions": false,
109
+ "output_hidden_states": false,
110
+ "output_scores": false,
111
+ "pad_token_id": null,
112
+ "prefix": null,
113
+ "problem_type": null,
114
+ "pruned_heads": {},
115
+ "qk_layer_norms_perceiver": false,
116
+ "remove_invalid_values": false,
117
+ "repetition_penalty": 1.0,
118
+ "resampler_depth": 6,
119
+ "resampler_head_dim": 96,
120
+ "resampler_n_heads": 16,
121
+ "resampler_n_latents": 64,
122
+ "return_dict": true,
123
+ "return_dict_in_generate": false,
124
+ "sep_token_id": null,
125
+ "suppress_tokens": null,
126
+ "task_specific_params": null,
127
+ "temperature": 1.0,
128
+ "tf_legacy_loss": false,
129
+ "tie_encoder_decoder": false,
130
+ "tie_word_embeddings": true,
131
+ "tokenizer_class": null,
132
+ "top_k": 50,
133
+ "top_p": 1.0,
134
+ "torch_dtype": null,
135
+ "torchscript": false,
136
+ "transformers_version": "4.46.0",
137
+ "typical_p": 1.0,
138
+ "use_bfloat16": false
139
+ },
140
+ "pixel_shuffle_factor": 4,
141
+ "pretraining_tp": 1,
142
+ "qk_layer_norms": false,
143
+ "rms_norm_eps": 1e-05,
144
+ "rope_interleaved": false,
145
+ "rope_scaling": null,
146
+ "rope_theta": 100000,
147
+ "transformers.js_config": {
148
+ "kv_cache_dtype": {
149
+ "fp16": "float16",
150
+ "q4f16": "float16"
151
+ }
152
+ },
153
+ "use_cache": true,
154
+ "use_resampler": false,
155
+ "vocab_size": 49280
156
+ },
157
+ "tie_word_embeddings": false,
158
+ "transformers.js_config": {
159
+ "kv_cache_dtype": {
160
+ "fp16": "float16",
161
+ "q4f16": "float16"
162
+ }
163
+ },
164
+ "use_cache": true,
165
+ "vision_config": {
166
+ "_attn_implementation_autoset": false,
167
+ "attention_dropout": 0.0,
168
+ "hidden_act": "gelu_pytorch_tanh",
169
+ "hidden_size": 768,
170
+ "image_size": 512,
171
+ "initializer_range": 0.02,
172
+ "intermediate_size": 3072,
173
+ "layer_norm_eps": 1e-06,
174
+ "max_image_size": {
175
+ "longest_edge": 512
176
+ },
177
+ "model_type": "idefics3_vision",
178
+ "num_attention_heads": 12,
179
+ "num_channels": 3,
180
+ "num_hidden_layers": 12,
181
+ "patch_size": 16,
182
+ "size": {
183
+ "longest_edge": 2048
184
+ },
185
+ "tie_word_embeddings": false,
186
+ "use_base_siglip": true
187
+ },
188
+ "vocab_size": 49280
189
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/2da7a00f0478d50ae1e7f75f085c5b2773b5f355f427c61cf34cb6febd629d96/7d7f3f7850d2aa89e3fa.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.6.dev4",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "prefill_chunk_size": 0,
40
+ "sequence_length": 1024,
41
+ "speculation_length": 0,
42
+ "start_rank_id": 0,
43
+ "target": "trn1",
44
+ "torch_dtype": "bfloat16",
45
+ "tp_degree": 2
46
+ },
47
+ "num_attention_heads": 2,
48
+ "num_hidden_layers": 2,
49
+ "num_key_value_heads": 1,
50
+ "original_max_position_embeddings": 16384,
51
+ "partial_rotary_factor": 1.0,
52
+ "resid_pdrop": 0.0,
53
+ "rms_norm_eps": 1e-05,
54
+ "rope_scaling": null,
55
+ "rope_theta": 250000,
56
+ "sliding_window": null,
57
+ "tie_word_embeddings": false,
58
+ "use_cache": true,
59
+ "vocab_size": 100352
60
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/441269935591cad8d370e512c0b93cdd2fce6247c40e5a4866d872ee5338b0de/01957724439347680e1b.json ADDED
@@ -0,0 +1,82 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "unsloth/gemma-3-270m-it",
4
+ "_sliding_window_pattern": 6,
5
+ "_task": "text-generation",
6
+ "architectures": [
7
+ "Gemma3ForCausalLM"
8
+ ],
9
+ "attention_bias": false,
10
+ "attention_dropout": 0.0,
11
+ "attn_logit_softcapping": null,
12
+ "dtype": "bfloat16",
13
+ "final_logit_softcapping": null,
14
+ "head_dim": 256,
15
+ "hidden_activation": "gelu_pytorch_tanh",
16
+ "hidden_size": 640,
17
+ "initializer_range": 0.02,
18
+ "intermediate_size": 2048,
19
+ "layer_types": [
20
+ "sliding_attention",
21
+ "sliding_attention",
22
+ "sliding_attention",
23
+ "sliding_attention",
24
+ "sliding_attention",
25
+ "full_attention",
26
+ "sliding_attention",
27
+ "sliding_attention",
28
+ "sliding_attention",
29
+ "sliding_attention",
30
+ "sliding_attention",
31
+ "full_attention",
32
+ "sliding_attention",
33
+ "sliding_attention",
34
+ "sliding_attention",
35
+ "sliding_attention",
36
+ "sliding_attention",
37
+ "full_attention"
38
+ ],
39
+ "max_position_embeddings": 32768,
40
+ "model_type": "gemma3_text",
41
+ "neuron": {
42
+ "_serialized_key": "NxDNeuronConfig",
43
+ "batch_size": 1,
44
+ "capacity_factor": null,
45
+ "checkpoint_id": "unsloth/gemma-3-270m-it",
46
+ "checkpoint_revision": "23cf460f6bb16954176b3ddcc8d4f250501458a9",
47
+ "continuous_batching": false,
48
+ "ep_degree": 1,
49
+ "fused_qkv": true,
50
+ "glu_mlp": true,
51
+ "local_ranks_size": 2,
52
+ "max_batch_size": 1,
53
+ "max_context_length": 1024,
54
+ "max_topk": 256,
55
+ "n_active_tokens": 1024,
56
+ "neuronxcc_version": "2.21.33363.0+82129205",
57
+ "on_device_sampling": true,
58
+ "optimum_neuron_version": "0.4.6.dev4",
59
+ "output_logits": false,
60
+ "pp_degree": 1,
61
+ "prefill_chunk_size": 0,
62
+ "sequence_length": 1024,
63
+ "speculation_length": 0,
64
+ "start_rank_id": 0,
65
+ "target": "trn1",
66
+ "torch_dtype": "bfloat16",
67
+ "tp_degree": 2
68
+ },
69
+ "num_attention_heads": 4,
70
+ "num_hidden_layers": 18,
71
+ "num_key_value_heads": 1,
72
+ "query_pre_attn_scalar": 256,
73
+ "rms_norm_eps": 1e-06,
74
+ "rope_local_base_freq": 10000.0,
75
+ "rope_scaling": null,
76
+ "rope_theta": 1000000.0,
77
+ "sliding_window": 512,
78
+ "unsloth_fixed": true,
79
+ "use_bidirectional_attention": false,
80
+ "use_cache": true,
81
+ "vocab_size": 262144
82
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/4ab8140bc7eb4a553d95855c5c2be2cf8c0fbab21b823d76183b6f51e98b6fc5/80038e075c3313d11a1d.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
33
+ "on_device_sampling": false,
34
+ "optimum_neuron_version": "0.4.6.dev4",
35
+ "output_logits": false,
36
+ "pp_degree": 1,
37
+ "prefill_chunk_size": 0,
38
+ "sequence_length": 1024,
39
+ "speculation_length": 0,
40
+ "start_rank_id": 0,
41
+ "target": "trn1",
42
+ "torch_dtype": "float16",
43
+ "tp_degree": 2
44
+ },
45
+ "num_attention_heads": 32,
46
+ "num_experts_per_tok": 2,
47
+ "num_hidden_layers": 2,
48
+ "num_key_value_heads": 8,
49
+ "num_local_experts": 8,
50
+ "output_router_logits": false,
51
+ "rms_norm_eps": 1e-05,
52
+ "rope_theta": 10000.0,
53
+ "router_aux_loss_coef": 0.001,
54
+ "router_jitter_noise": 0.0,
55
+ "sliding_window": 4096,
56
+ "tie_word_embeddings": false,
57
+ "use_cache": true,
58
+ "vocab_size": 32000
59
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/4cb7aff9e2a15c151396f2b684013e39d6739f0dec83e5c9dabbfe9d5fcf77b7/e28a26007281cdef7599.json ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen2.5-0.5B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen2ForCausalLM"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "dtype": "bfloat16",
10
+ "hidden_act": "silu",
11
+ "hidden_size": 896,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 4864,
14
+ "layer_types": [
15
+ "full_attention",
16
+ "full_attention",
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
+ ],
40
+ "max_position_embeddings": 32768,
41
+ "max_window_layers": 24,
42
+ "model_type": "qwen2",
43
+ "neuron": {
44
+ "_serialized_key": "NxDNeuronConfig",
45
+ "batch_size": 2,
46
+ "capacity_factor": null,
47
+ "checkpoint_id": "Qwen/Qwen2.5-0.5B",
48
+ "checkpoint_revision": "060db6499f32faf8b98477b0a26969ef7d8b9987",
49
+ "continuous_batching": true,
50
+ "ep_degree": 1,
51
+ "fused_qkv": false,
52
+ "glu_mlp": true,
53
+ "local_ranks_size": 2,
54
+ "max_batch_size": 2,
55
+ "max_context_length": 1024,
56
+ "max_topk": 256,
57
+ "n_active_tokens": 1024,
58
+ "neuronxcc_version": "2.21.33363.0+82129205",
59
+ "on_device_sampling": false,
60
+ "optimum_neuron_version": "0.4.6.dev4",
61
+ "output_logits": false,
62
+ "pp_degree": 1,
63
+ "prefill_chunk_size": 0,
64
+ "sequence_length": 1024,
65
+ "speculation_length": 0,
66
+ "start_rank_id": 0,
67
+ "target": "trn1",
68
+ "torch_dtype": "bfloat16",
69
+ "tp_degree": 2
70
+ },
71
+ "num_attention_heads": 14,
72
+ "num_hidden_layers": 24,
73
+ "num_key_value_heads": 2,
74
+ "rms_norm_eps": 1e-06,
75
+ "rope_scaling": null,
76
+ "rope_theta": 1000000.0,
77
+ "sliding_window": null,
78
+ "tie_word_embeddings": true,
79
+ "use_cache": true,
80
+ "use_mrope": false,
81
+ "use_sliding_window": false,
82
+ "vocab_size": 151936
83
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/6454afdf3e9d66c7226c13a575b718845c25e53b0699600ba2bb4f883e9d841b/b2922eb58f0f3319bfe3.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.6.dev4",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "prefill_chunk_size": 0,
40
+ "sequence_length": 1024,
41
+ "speculation_length": 0,
42
+ "start_rank_id": 0,
43
+ "target": "trn1",
44
+ "torch_dtype": "float16",
45
+ "tp_degree": 2
46
+ },
47
+ "num_attention_heads": 4,
48
+ "num_hidden_layers": 2,
49
+ "num_key_value_heads": 4,
50
+ "pretraining_tp": 1,
51
+ "rms_norm_eps": 1e-05,
52
+ "rope_scaling": {
53
+ "factor": 8.0,
54
+ "high_freq_factor": 4.0,
55
+ "low_freq_factor": 1.0,
56
+ "original_max_position_embeddings": 8192,
57
+ "rope_type": "llama3"
58
+ },
59
+ "rope_theta": 500000.0,
60
+ "tie_word_embeddings": false,
61
+ "use_cache": true,
62
+ "vocab_size": 128256
63
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/73707b485eab9008c7aba7f5dad0ce2384ac685318d5f888c12fa0d81ed90b19/3aed09186384cc6540ef.json ADDED
@@ -0,0 +1,135 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "HuggingFaceTB/SmolLM3-3B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "SmolLM3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "hidden_act": "silu",
12
+ "hidden_size": 2048,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 11008,
15
+ "layer_types": [
16
+ "full_attention",
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
+ ],
53
+ "max_position_embeddings": 65536,
54
+ "max_window_layers": 28,
55
+ "mlp_bias": false,
56
+ "model_type": "smollm3",
57
+ "neuron": {
58
+ "_serialized_key": "NxDNeuronConfig",
59
+ "batch_size": 1,
60
+ "capacity_factor": null,
61
+ "checkpoint_id": "HuggingFaceTB/SmolLM3-3B",
62
+ "checkpoint_revision": "a07cc9a04f16550a088caea529712d1d335b0ac1",
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": 1024,
70
+ "max_topk": 256,
71
+ "n_active_tokens": 1024,
72
+ "neuronxcc_version": "2.21.33363.0+82129205",
73
+ "on_device_sampling": true,
74
+ "optimum_neuron_version": "0.4.6.dev4",
75
+ "output_logits": false,
76
+ "pp_degree": 1,
77
+ "prefill_chunk_size": 0,
78
+ "sequence_length": 1024,
79
+ "speculation_length": 0,
80
+ "start_rank_id": 0,
81
+ "target": "trn1",
82
+ "torch_dtype": "bfloat16",
83
+ "tp_degree": 2
84
+ },
85
+ "no_rope_layer_interval": 4,
86
+ "no_rope_layers": [
87
+ 1,
88
+ 1,
89
+ 1,
90
+ 0,
91
+ 1,
92
+ 1,
93
+ 1,
94
+ 0,
95
+ 1,
96
+ 1,
97
+ 1,
98
+ 0,
99
+ 1,
100
+ 1,
101
+ 1,
102
+ 0,
103
+ 1,
104
+ 1,
105
+ 1,
106
+ 0,
107
+ 1,
108
+ 1,
109
+ 1,
110
+ 0,
111
+ 1,
112
+ 1,
113
+ 1,
114
+ 0,
115
+ 1,
116
+ 1,
117
+ 1,
118
+ 0,
119
+ 1,
120
+ 1,
121
+ 1,
122
+ 0
123
+ ],
124
+ "num_attention_heads": 16,
125
+ "num_hidden_layers": 36,
126
+ "num_key_value_heads": 4,
127
+ "pretraining_tp": 2,
128
+ "rms_norm_eps": 1e-06,
129
+ "rope_scaling": null,
130
+ "rope_theta": 5000000.0,
131
+ "sliding_window": null,
132
+ "use_cache": false,
133
+ "use_sliding_window": false,
134
+ "vocab_size": 128256
135
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/7518518c7e077820070186deda960d8cc49db068cdf0ac70664098fa2b6b698c/c7e74a139ce290675aa6.json ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.6.dev4",
39
+ "output_logits": false,
40
+ "pp_degree": 1,
41
+ "prefill_chunk_size": 0,
42
+ "sequence_length": 1024,
43
+ "speculation_length": 0,
44
+ "start_rank_id": 0,
45
+ "target": "trn1",
46
+ "torch_dtype": "bfloat16",
47
+ "tp_degree": 2
48
+ },
49
+ "num_attention_heads": 4,
50
+ "num_hidden_layers": 2,
51
+ "num_key_value_heads": 2,
52
+ "rms_norm_eps": 1e-06,
53
+ "rope_scaling": {
54
+ "factor": 4.0,
55
+ "original_max_position_embeddings": 32768,
56
+ "rope_type": "yarn",
57
+ "type": "yarn"
58
+ },
59
+ "rope_theta": 1000000.0,
60
+ "sliding_window": null,
61
+ "tie_word_embeddings": false,
62
+ "use_cache": true,
63
+ "use_sliding_window": false,
64
+ "vocab_size": 152064
65
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/7f05bde17c7b0ffeb657897697f23d182f406b76ced7f1b2cd5741dc93fe2e2e/ca71d7a47e38a69da803.json ADDED
@@ -0,0 +1,126 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "tiny-random/llama-4",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Llama4ForConditionalGeneration"
7
+ ],
8
+ "boi_token_index": 200080,
9
+ "dtype": "bfloat16",
10
+ "eoi_token_index": 200081,
11
+ "image_token_index": 200092,
12
+ "model_type": "llama4",
13
+ "neuron": {
14
+ "_serialized_key": "NxDNeuronConfig",
15
+ "batch_size": 1,
16
+ "capacity_factor": null,
17
+ "checkpoint_id": "tiny-random/llama-4",
18
+ "checkpoint_revision": "9e716f5d4d1ffe0a44a15f46f4a12b840439aba4",
19
+ "continuous_batching": false,
20
+ "ep_degree": 1,
21
+ "fused_qkv": false,
22
+ "glu_mlp": true,
23
+ "local_ranks_size": 2,
24
+ "max_batch_size": 1,
25
+ "max_context_length": 1024,
26
+ "max_topk": 256,
27
+ "n_active_tokens": 1024,
28
+ "neuronxcc_version": "2.21.33363.0+82129205",
29
+ "on_device_sampling": true,
30
+ "optimum_neuron_version": "0.4.6.dev4",
31
+ "output_logits": false,
32
+ "pp_degree": 1,
33
+ "prefill_chunk_size": 0,
34
+ "sequence_length": 1024,
35
+ "speculation_length": 0,
36
+ "start_rank_id": 0,
37
+ "target": "trn1",
38
+ "torch_dtype": "bfloat16",
39
+ "tp_degree": 2
40
+ },
41
+ "text_config": {
42
+ "_attn_implementation_autoset": true,
43
+ "attention_bias": false,
44
+ "attention_chunk_size": 128,
45
+ "attention_dropout": 0.0,
46
+ "attn_scale": 0.1,
47
+ "attn_temperature_tuning": 4,
48
+ "bos_token_id": 200000,
49
+ "cache_implementation": "hybrid",
50
+ "dtype": "bfloat16",
51
+ "eos_token_id": [
52
+ 200001,
53
+ 200007,
54
+ 200008
55
+ ],
56
+ "floor_scale": 8192,
57
+ "for_llm_compressor": false,
58
+ "head_dim": 32,
59
+ "hidden_act": "silu",
60
+ "hidden_size": 32,
61
+ "initializer_range": 0.02,
62
+ "interleave_moe_layer_step": 2,
63
+ "intermediate_size": 64,
64
+ "intermediate_size_mlp": 128,
65
+ "layer_types": [
66
+ "chunked_attention",
67
+ "chunked_attention",
68
+ "chunked_attention",
69
+ "full_attention"
70
+ ],
71
+ "max_position_embeddings": 1048576,
72
+ "model_type": "llama4_text",
73
+ "moe_layers": [
74
+ 1,
75
+ 3
76
+ ],
77
+ "no_rope_layers": [
78
+ 1,
79
+ 1,
80
+ 1,
81
+ 0
82
+ ],
83
+ "num_attention_heads": 1,
84
+ "num_experts_per_tok": 1,
85
+ "num_hidden_layers": 4,
86
+ "num_key_value_heads": 1,
87
+ "num_local_experts": 8,
88
+ "output_router_logits": false,
89
+ "pad_token_id": 200018,
90
+ "rms_norm_eps": 1e-05,
91
+ "rope_scaling": null,
92
+ "rope_theta": 500000.0,
93
+ "router_aux_loss_coef": 0.001,
94
+ "router_jitter_noise": 0.0,
95
+ "tie_word_embeddings": true,
96
+ "use_cache": true,
97
+ "use_qk_norm": true,
98
+ "vocab_size": 202048
99
+ },
100
+ "tie_word_embeddings": false,
101
+ "vision_config": {
102
+ "_attn_implementation_autoset": true,
103
+ "_vision_feature_layer": -1,
104
+ "attention_dropout": 0.0,
105
+ "hidden_act": "gelu",
106
+ "hidden_size": 32,
107
+ "image_size": 336,
108
+ "initializer_range": 0.02,
109
+ "intermediate_size": 128,
110
+ "model_type": "llama4_vision_model",
111
+ "multi_modal_projector_bias": false,
112
+ "norm_eps": 1e-05,
113
+ "num_attention_heads": 1,
114
+ "num_channels": 3,
115
+ "num_hidden_layers": 2,
116
+ "patch_size": 14,
117
+ "pixel_shuffle_ratio": 0.5,
118
+ "projector_dropout": 0.0,
119
+ "projector_input_dim": 32,
120
+ "projector_output_dim": 32,
121
+ "rope_theta": 10000,
122
+ "vision_feature_layer": -1,
123
+ "vision_feature_select_strategy": "default",
124
+ "vision_output_dim": 32
125
+ }
126
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/920f44ce6d3e004d1ce547ae06644f7be262180644b04573153aa15d98742edc/d6743412c8cf7830ae4a.json ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
38
+ "on_device_sampling": true,
39
+ "optimum_neuron_version": "0.4.6.dev4",
40
+ "output_logits": false,
41
+ "pp_degree": 1,
42
+ "prefill_chunk_size": 0,
43
+ "sequence_length": 1024,
44
+ "speculation_length": 0,
45
+ "start_rank_id": 0,
46
+ "target": "trn1",
47
+ "torch_dtype": "float32",
48
+ "tp_degree": 2
49
+ },
50
+ "norm_topk_prob": true,
51
+ "num_attention_heads": 2,
52
+ "num_experts": 8,
53
+ "num_experts_per_tok": 2,
54
+ "num_hidden_layers": 2,
55
+ "num_key_value_heads": 1,
56
+ "output_router_logits": false,
57
+ "rms_norm_eps": 1e-06,
58
+ "rope_scaling": null,
59
+ "rope_theta": 1000000.0,
60
+ "router_aux_loss_coef": 0.001,
61
+ "sliding_window": null,
62
+ "tie_word_embeddings": true,
63
+ "use_cache": true,
64
+ "use_sliding_window": false,
65
+ "vocab_size": 151936
66
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/cf6b9a360dcf294104671106bae2adbd9fd291823bb60a351883163684073231/7d57cf017fab771c9e95.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "unsloth/Llama-3.2-1B-Instruct",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "LlamaForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 64,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 2048,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 8192,
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": "unsloth/Llama-3.2-1B-Instruct",
24
+ "checkpoint_revision": "5a8abab4a5d6f164389b1079fb721cfab8d7126c",
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": 4096,
32
+ "max_topk": 256,
33
+ "n_active_tokens": 4096,
34
+ "neuronxcc_version": "2.21.33363.0+82129205",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.6.dev4",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "prefill_chunk_size": 1024,
40
+ "sequence_length": 4096,
41
+ "speculation_length": 0,
42
+ "start_rank_id": 0,
43
+ "target": "trn1",
44
+ "torch_dtype": "bfloat16",
45
+ "tp_degree": 2
46
+ },
47
+ "num_attention_heads": 32,
48
+ "num_hidden_layers": 16,
49
+ "num_key_value_heads": 8,
50
+ "pretraining_tp": 1,
51
+ "rms_norm_eps": 1e-05,
52
+ "rope_scaling": {
53
+ "factor": 32.0,
54
+ "high_freq_factor": 4.0,
55
+ "low_freq_factor": 1.0,
56
+ "original_max_position_embeddings": 8192,
57
+ "rope_type": "llama3"
58
+ },
59
+ "rope_theta": 500000.0,
60
+ "tie_word_embeddings": true,
61
+ "unsloth_fixed": true,
62
+ "use_cache": true,
63
+ "vocab_size": 128256
64
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/cf6b9a360dcf294104671106bae2adbd9fd291823bb60a351883163684073231/9fdf2fe50b15951e2b95.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "unsloth/Llama-3.2-1B-Instruct",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "LlamaForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 64,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 2048,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 8192,
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": "unsloth/Llama-3.2-1B-Instruct",
24
+ "checkpoint_revision": null,
25
+ "continuous_batching": false,
26
+ "ep_degree": 1,
27
+ "fused_qkv": false,
28
+ "glu_mlp": true,
29
+ "local_ranks_size": 2,
30
+ "max_batch_size": 1,
31
+ "max_context_length": 4096,
32
+ "max_topk": 256,
33
+ "n_active_tokens": 4096,
34
+ "neuronxcc_version": "2.21.33363.0+82129205",
35
+ "on_device_sampling": false,
36
+ "optimum_neuron_version": "0.4.6.dev4",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "prefill_chunk_size": 0,
40
+ "sequence_length": 4096,
41
+ "speculation_length": 0,
42
+ "start_rank_id": 0,
43
+ "target": "trn1",
44
+ "torch_dtype": "bfloat16",
45
+ "tp_degree": 2
46
+ },
47
+ "num_attention_heads": 32,
48
+ "num_hidden_layers": 16,
49
+ "num_key_value_heads": 8,
50
+ "pretraining_tp": 1,
51
+ "rms_norm_eps": 1e-05,
52
+ "rope_scaling": {
53
+ "factor": 32.0,
54
+ "high_freq_factor": 4.0,
55
+ "low_freq_factor": 1.0,
56
+ "original_max_position_embeddings": 8192,
57
+ "rope_type": "llama3"
58
+ },
59
+ "rope_theta": 500000.0,
60
+ "tie_word_embeddings": true,
61
+ "unsloth_fixed": true,
62
+ "use_cache": true,
63
+ "vocab_size": 128256
64
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/cf6b9a360dcf294104671106bae2adbd9fd291823bb60a351883163684073231/c737e0d82a12fca4c663.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "unsloth/Llama-3.2-1B-Instruct",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "LlamaForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 64,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 2048,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 8192,
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": "unsloth/Llama-3.2-1B-Instruct",
24
+ "checkpoint_revision": null,
25
+ "continuous_batching": false,
26
+ "ep_degree": 1,
27
+ "fused_qkv": false,
28
+ "glu_mlp": true,
29
+ "local_ranks_size": 2,
30
+ "max_batch_size": 1,
31
+ "max_context_length": 4096,
32
+ "max_topk": 256,
33
+ "n_active_tokens": 4096,
34
+ "neuronxcc_version": "2.21.33363.0+82129205",
35
+ "on_device_sampling": false,
36
+ "optimum_neuron_version": "0.4.6.dev4",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "prefill_chunk_size": 0,
40
+ "sequence_length": 4096,
41
+ "speculation_length": 5,
42
+ "start_rank_id": 0,
43
+ "target": "trn1",
44
+ "torch_dtype": "bfloat16",
45
+ "tp_degree": 2
46
+ },
47
+ "num_attention_heads": 32,
48
+ "num_hidden_layers": 16,
49
+ "num_key_value_heads": 8,
50
+ "pretraining_tp": 1,
51
+ "rms_norm_eps": 1e-05,
52
+ "rope_scaling": {
53
+ "factor": 32.0,
54
+ "high_freq_factor": 4.0,
55
+ "low_freq_factor": 1.0,
56
+ "original_max_position_embeddings": 8192,
57
+ "rope_type": "llama3"
58
+ },
59
+ "rope_theta": 500000.0,
60
+ "tie_word_embeddings": true,
61
+ "unsloth_fixed": true,
62
+ "use_cache": true,
63
+ "vocab_size": 128256
64
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/d139acf64685f15794bb983ff6eb881bdd31304bae88b0ce1ed20a54c21f2265/fcca5755ba87a7768913.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.6.dev4",
39
+ "output_logits": false,
40
+ "pp_degree": 1,
41
+ "prefill_chunk_size": 0,
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
+ "num_attention_heads": 4,
50
+ "num_hidden_layers": 2,
51
+ "num_key_value_heads": 4,
52
+ "residual_multiplier": 1.0,
53
+ "rms_norm_eps": 1e-06,
54
+ "rope_scaling": null,
55
+ "rope_theta": 10000.0,
56
+ "tie_word_embeddings": false,
57
+ "use_cache": true,
58
+ "vocab_size": 49152
59
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/gemma3_text/unsloth/gemma-3-270m-it/01957724439347680e1b.json ADDED
@@ -0,0 +1,82 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "unsloth/gemma-3-270m-it",
4
+ "_sliding_window_pattern": 6,
5
+ "_task": "text-generation",
6
+ "architectures": [
7
+ "Gemma3ForCausalLM"
8
+ ],
9
+ "attention_bias": false,
10
+ "attention_dropout": 0.0,
11
+ "attn_logit_softcapping": null,
12
+ "dtype": "bfloat16",
13
+ "final_logit_softcapping": null,
14
+ "head_dim": 256,
15
+ "hidden_activation": "gelu_pytorch_tanh",
16
+ "hidden_size": 640,
17
+ "initializer_range": 0.02,
18
+ "intermediate_size": 2048,
19
+ "layer_types": [
20
+ "sliding_attention",
21
+ "sliding_attention",
22
+ "sliding_attention",
23
+ "sliding_attention",
24
+ "sliding_attention",
25
+ "full_attention",
26
+ "sliding_attention",
27
+ "sliding_attention",
28
+ "sliding_attention",
29
+ "sliding_attention",
30
+ "sliding_attention",
31
+ "full_attention",
32
+ "sliding_attention",
33
+ "sliding_attention",
34
+ "sliding_attention",
35
+ "sliding_attention",
36
+ "sliding_attention",
37
+ "full_attention"
38
+ ],
39
+ "max_position_embeddings": 32768,
40
+ "model_type": "gemma3_text",
41
+ "neuron": {
42
+ "_serialized_key": "NxDNeuronConfig",
43
+ "batch_size": 1,
44
+ "capacity_factor": null,
45
+ "checkpoint_id": "unsloth/gemma-3-270m-it",
46
+ "checkpoint_revision": "23cf460f6bb16954176b3ddcc8d4f250501458a9",
47
+ "continuous_batching": false,
48
+ "ep_degree": 1,
49
+ "fused_qkv": true,
50
+ "glu_mlp": true,
51
+ "local_ranks_size": 2,
52
+ "max_batch_size": 1,
53
+ "max_context_length": 1024,
54
+ "max_topk": 256,
55
+ "n_active_tokens": 1024,
56
+ "neuronxcc_version": "2.21.33363.0+82129205",
57
+ "on_device_sampling": true,
58
+ "optimum_neuron_version": "0.4.6.dev4",
59
+ "output_logits": false,
60
+ "pp_degree": 1,
61
+ "prefill_chunk_size": 0,
62
+ "sequence_length": 1024,
63
+ "speculation_length": 0,
64
+ "start_rank_id": 0,
65
+ "target": "trn1",
66
+ "torch_dtype": "bfloat16",
67
+ "tp_degree": 2
68
+ },
69
+ "num_attention_heads": 4,
70
+ "num_hidden_layers": 18,
71
+ "num_key_value_heads": 1,
72
+ "query_pre_attn_scalar": 256,
73
+ "rms_norm_eps": 1e-06,
74
+ "rope_local_base_freq": 10000.0,
75
+ "rope_scaling": null,
76
+ "rope_theta": 1000000.0,
77
+ "sliding_window": 512,
78
+ "unsloth_fixed": true,
79
+ "use_bidirectional_attention": false,
80
+ "use_cache": true,
81
+ "vocab_size": 262144
82
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/fcca5755ba87a7768913.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.6.dev4",
39
+ "output_logits": false,
40
+ "pp_degree": 1,
41
+ "prefill_chunk_size": 0,
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
+ "num_attention_heads": 4,
50
+ "num_hidden_layers": 2,
51
+ "num_key_value_heads": 4,
52
+ "residual_multiplier": 1.0,
53
+ "rms_norm_eps": 1e-06,
54
+ "rope_scaling": null,
55
+ "rope_theta": 10000.0,
56
+ "tie_word_embeddings": false,
57
+ "use_cache": true,
58
+ "vocab_size": 49152
59
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/idefics3/HuggingFaceTB/SmolVLM-256M-Instruct/2339cde11cbd17accc62.json ADDED
@@ -0,0 +1,189 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "HuggingFaceTB/SmolVLM-256M-Instruct",
4
+ "_task": "image-text-to-text",
5
+ "architectures": [
6
+ "Idefics3ForConditionalGeneration"
7
+ ],
8
+ "dtype": "bfloat16",
9
+ "image_token_id": 49190,
10
+ "model_type": "idefics3",
11
+ "neuron": {
12
+ "_serialized_key": "NxDVLMNeuronConfig",
13
+ "batch_size": 1,
14
+ "capacity_factor": null,
15
+ "checkpoint_id": "HuggingFaceTB/SmolVLM-256M-Instruct",
16
+ "checkpoint_revision": "7e3e67edbbed1bf9888184d9df282b700a323964",
17
+ "continuous_batching": false,
18
+ "ep_degree": 1,
19
+ "fused_qkv": true,
20
+ "glu_mlp": true,
21
+ "image_seq_len": 64,
22
+ "image_size": 512,
23
+ "local_ranks_size": 2,
24
+ "max_batch_size": 1,
25
+ "max_context_length": 2048,
26
+ "max_num_images": 17,
27
+ "max_topk": 256,
28
+ "n_active_tokens": 2048,
29
+ "neuronxcc_version": "2.21.33363.0+82129205",
30
+ "on_device_sampling": true,
31
+ "optimum_neuron_version": "0.4.6.dev4",
32
+ "output_logits": false,
33
+ "pp_degree": 1,
34
+ "prefill_chunk_size": 1024,
35
+ "sequence_length": 2048,
36
+ "speculation_length": 0,
37
+ "start_rank_id": 0,
38
+ "target": "trn1",
39
+ "torch_dtype": "bfloat16",
40
+ "tp_degree": 2
41
+ },
42
+ "scale_factor": 4,
43
+ "text_config": {
44
+ "_attn_implementation_autoset": false,
45
+ "_flash_attn_2_enabled": true,
46
+ "_name_or_path": "None",
47
+ "architectures": [
48
+ "VLlama3ForCausalLM"
49
+ ],
50
+ "attention_bias": false,
51
+ "attention_dropout": 0.0,
52
+ "dtype": "bfloat16",
53
+ "head_dim": 64,
54
+ "hidden_act": "silu",
55
+ "hidden_size": 576,
56
+ "initializer_range": 0.041666666666666664,
57
+ "intermediate_size": 1536,
58
+ "is_llama_config": true,
59
+ "max_position_embeddings": 8192,
60
+ "mlp_bias": false,
61
+ "model_type": "llama",
62
+ "neftune_noise_alpha": 0.0,
63
+ "num_attention_heads": 9,
64
+ "num_hidden_layers": 30,
65
+ "num_key_value_heads": 3,
66
+ "pad_token_id": 2,
67
+ "perceiver_config": {
68
+ "_attn_implementation_autoset": false,
69
+ "_name_or_path": "",
70
+ "add_cross_attention": false,
71
+ "architectures": null,
72
+ "attention_dropout": 0.0,
73
+ "bad_words_ids": null,
74
+ "begin_suppress_tokens": null,
75
+ "bos_token_id": null,
76
+ "chunk_size_feed_forward": 0,
77
+ "cross_attention_hidden_size": null,
78
+ "decoder_start_token_id": null,
79
+ "diversity_penalty": 0.0,
80
+ "do_sample": false,
81
+ "early_stopping": false,
82
+ "encoder_no_repeat_ngram_size": 0,
83
+ "eos_token_id": null,
84
+ "exponential_decay_length_penalty": null,
85
+ "finetuning_task": null,
86
+ "forced_bos_token_id": null,
87
+ "forced_eos_token_id": null,
88
+ "hidden_act": "silu",
89
+ "id2label": {
90
+ "0": "LABEL_0",
91
+ "1": "LABEL_1"
92
+ },
93
+ "is_decoder": false,
94
+ "is_encoder_decoder": false,
95
+ "label2id": {
96
+ "LABEL_0": 0,
97
+ "LABEL_1": 1
98
+ },
99
+ "length_penalty": 1.0,
100
+ "max_length": 20,
101
+ "min_length": 0,
102
+ "model_type": "vllama3",
103
+ "no_repeat_ngram_size": 0,
104
+ "num_beam_groups": 1,
105
+ "num_beams": 1,
106
+ "num_key_value_heads": 1,
107
+ "num_return_sequences": 1,
108
+ "output_attentions": false,
109
+ "output_hidden_states": false,
110
+ "output_scores": false,
111
+ "pad_token_id": null,
112
+ "prefix": null,
113
+ "problem_type": null,
114
+ "pruned_heads": {},
115
+ "qk_layer_norms_perceiver": false,
116
+ "remove_invalid_values": false,
117
+ "repetition_penalty": 1.0,
118
+ "resampler_depth": 6,
119
+ "resampler_head_dim": 96,
120
+ "resampler_n_heads": 16,
121
+ "resampler_n_latents": 64,
122
+ "return_dict": true,
123
+ "return_dict_in_generate": false,
124
+ "sep_token_id": null,
125
+ "suppress_tokens": null,
126
+ "task_specific_params": null,
127
+ "temperature": 1.0,
128
+ "tf_legacy_loss": false,
129
+ "tie_encoder_decoder": false,
130
+ "tie_word_embeddings": true,
131
+ "tokenizer_class": null,
132
+ "top_k": 50,
133
+ "top_p": 1.0,
134
+ "torch_dtype": null,
135
+ "torchscript": false,
136
+ "transformers_version": "4.46.0",
137
+ "typical_p": 1.0,
138
+ "use_bfloat16": false
139
+ },
140
+ "pixel_shuffle_factor": 4,
141
+ "pretraining_tp": 1,
142
+ "qk_layer_norms": false,
143
+ "rms_norm_eps": 1e-05,
144
+ "rope_interleaved": false,
145
+ "rope_scaling": null,
146
+ "rope_theta": 100000,
147
+ "transformers.js_config": {
148
+ "kv_cache_dtype": {
149
+ "fp16": "float16",
150
+ "q4f16": "float16"
151
+ }
152
+ },
153
+ "use_cache": true,
154
+ "use_resampler": false,
155
+ "vocab_size": 49280
156
+ },
157
+ "tie_word_embeddings": false,
158
+ "transformers.js_config": {
159
+ "kv_cache_dtype": {
160
+ "fp16": "float16",
161
+ "q4f16": "float16"
162
+ }
163
+ },
164
+ "use_cache": true,
165
+ "vision_config": {
166
+ "_attn_implementation_autoset": false,
167
+ "attention_dropout": 0.0,
168
+ "hidden_act": "gelu_pytorch_tanh",
169
+ "hidden_size": 768,
170
+ "image_size": 512,
171
+ "initializer_range": 0.02,
172
+ "intermediate_size": 3072,
173
+ "layer_norm_eps": 1e-06,
174
+ "max_image_size": {
175
+ "longest_edge": 512
176
+ },
177
+ "model_type": "idefics3_vision",
178
+ "num_attention_heads": 12,
179
+ "num_channels": 3,
180
+ "num_hidden_layers": 12,
181
+ "patch_size": 16,
182
+ "size": {
183
+ "longest_edge": 2048
184
+ },
185
+ "tie_word_embeddings": false,
186
+ "use_base_siglip": true
187
+ },
188
+ "vocab_size": 49280
189
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/llama/llamafactory/tiny-random-Llama-3/b2922eb58f0f3319bfe3.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.6.dev4",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "prefill_chunk_size": 0,
40
+ "sequence_length": 1024,
41
+ "speculation_length": 0,
42
+ "start_rank_id": 0,
43
+ "target": "trn1",
44
+ "torch_dtype": "float16",
45
+ "tp_degree": 2
46
+ },
47
+ "num_attention_heads": 4,
48
+ "num_hidden_layers": 2,
49
+ "num_key_value_heads": 4,
50
+ "pretraining_tp": 1,
51
+ "rms_norm_eps": 1e-05,
52
+ "rope_scaling": {
53
+ "factor": 8.0,
54
+ "high_freq_factor": 4.0,
55
+ "low_freq_factor": 1.0,
56
+ "original_max_position_embeddings": 8192,
57
+ "rope_type": "llama3"
58
+ },
59
+ "rope_theta": 500000.0,
60
+ "tie_word_embeddings": false,
61
+ "use_cache": true,
62
+ "vocab_size": 128256
63
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/llama/unsloth/Llama-3.2-1B-Instruct/c737e0d82a12fca4c663.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "unsloth/Llama-3.2-1B-Instruct",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "LlamaForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 64,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 2048,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 8192,
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": "unsloth/Llama-3.2-1B-Instruct",
24
+ "checkpoint_revision": null,
25
+ "continuous_batching": false,
26
+ "ep_degree": 1,
27
+ "fused_qkv": false,
28
+ "glu_mlp": true,
29
+ "local_ranks_size": 2,
30
+ "max_batch_size": 1,
31
+ "max_context_length": 4096,
32
+ "max_topk": 256,
33
+ "n_active_tokens": 4096,
34
+ "neuronxcc_version": "2.21.33363.0+82129205",
35
+ "on_device_sampling": false,
36
+ "optimum_neuron_version": "0.4.6.dev4",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "prefill_chunk_size": 0,
40
+ "sequence_length": 4096,
41
+ "speculation_length": 5,
42
+ "start_rank_id": 0,
43
+ "target": "trn1",
44
+ "torch_dtype": "bfloat16",
45
+ "tp_degree": 2
46
+ },
47
+ "num_attention_heads": 32,
48
+ "num_hidden_layers": 16,
49
+ "num_key_value_heads": 8,
50
+ "pretraining_tp": 1,
51
+ "rms_norm_eps": 1e-05,
52
+ "rope_scaling": {
53
+ "factor": 32.0,
54
+ "high_freq_factor": 4.0,
55
+ "low_freq_factor": 1.0,
56
+ "original_max_position_embeddings": 8192,
57
+ "rope_type": "llama3"
58
+ },
59
+ "rope_theta": 500000.0,
60
+ "tie_word_embeddings": true,
61
+ "unsloth_fixed": true,
62
+ "use_cache": true,
63
+ "vocab_size": 128256
64
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/llama4/tiny-random/llama-4/ca71d7a47e38a69da803.json ADDED
@@ -0,0 +1,126 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "tiny-random/llama-4",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Llama4ForConditionalGeneration"
7
+ ],
8
+ "boi_token_index": 200080,
9
+ "dtype": "bfloat16",
10
+ "eoi_token_index": 200081,
11
+ "image_token_index": 200092,
12
+ "model_type": "llama4",
13
+ "neuron": {
14
+ "_serialized_key": "NxDNeuronConfig",
15
+ "batch_size": 1,
16
+ "capacity_factor": null,
17
+ "checkpoint_id": "tiny-random/llama-4",
18
+ "checkpoint_revision": "9e716f5d4d1ffe0a44a15f46f4a12b840439aba4",
19
+ "continuous_batching": false,
20
+ "ep_degree": 1,
21
+ "fused_qkv": false,
22
+ "glu_mlp": true,
23
+ "local_ranks_size": 2,
24
+ "max_batch_size": 1,
25
+ "max_context_length": 1024,
26
+ "max_topk": 256,
27
+ "n_active_tokens": 1024,
28
+ "neuronxcc_version": "2.21.33363.0+82129205",
29
+ "on_device_sampling": true,
30
+ "optimum_neuron_version": "0.4.6.dev4",
31
+ "output_logits": false,
32
+ "pp_degree": 1,
33
+ "prefill_chunk_size": 0,
34
+ "sequence_length": 1024,
35
+ "speculation_length": 0,
36
+ "start_rank_id": 0,
37
+ "target": "trn1",
38
+ "torch_dtype": "bfloat16",
39
+ "tp_degree": 2
40
+ },
41
+ "text_config": {
42
+ "_attn_implementation_autoset": true,
43
+ "attention_bias": false,
44
+ "attention_chunk_size": 128,
45
+ "attention_dropout": 0.0,
46
+ "attn_scale": 0.1,
47
+ "attn_temperature_tuning": 4,
48
+ "bos_token_id": 200000,
49
+ "cache_implementation": "hybrid",
50
+ "dtype": "bfloat16",
51
+ "eos_token_id": [
52
+ 200001,
53
+ 200007,
54
+ 200008
55
+ ],
56
+ "floor_scale": 8192,
57
+ "for_llm_compressor": false,
58
+ "head_dim": 32,
59
+ "hidden_act": "silu",
60
+ "hidden_size": 32,
61
+ "initializer_range": 0.02,
62
+ "interleave_moe_layer_step": 2,
63
+ "intermediate_size": 64,
64
+ "intermediate_size_mlp": 128,
65
+ "layer_types": [
66
+ "chunked_attention",
67
+ "chunked_attention",
68
+ "chunked_attention",
69
+ "full_attention"
70
+ ],
71
+ "max_position_embeddings": 1048576,
72
+ "model_type": "llama4_text",
73
+ "moe_layers": [
74
+ 1,
75
+ 3
76
+ ],
77
+ "no_rope_layers": [
78
+ 1,
79
+ 1,
80
+ 1,
81
+ 0
82
+ ],
83
+ "num_attention_heads": 1,
84
+ "num_experts_per_tok": 1,
85
+ "num_hidden_layers": 4,
86
+ "num_key_value_heads": 1,
87
+ "num_local_experts": 8,
88
+ "output_router_logits": false,
89
+ "pad_token_id": 200018,
90
+ "rms_norm_eps": 1e-05,
91
+ "rope_scaling": null,
92
+ "rope_theta": 500000.0,
93
+ "router_aux_loss_coef": 0.001,
94
+ "router_jitter_noise": 0.0,
95
+ "tie_word_embeddings": true,
96
+ "use_cache": true,
97
+ "use_qk_norm": true,
98
+ "vocab_size": 202048
99
+ },
100
+ "tie_word_embeddings": false,
101
+ "vision_config": {
102
+ "_attn_implementation_autoset": true,
103
+ "_vision_feature_layer": -1,
104
+ "attention_dropout": 0.0,
105
+ "hidden_act": "gelu",
106
+ "hidden_size": 32,
107
+ "image_size": 336,
108
+ "initializer_range": 0.02,
109
+ "intermediate_size": 128,
110
+ "model_type": "llama4_vision_model",
111
+ "multi_modal_projector_bias": false,
112
+ "norm_eps": 1e-05,
113
+ "num_attention_heads": 1,
114
+ "num_channels": 3,
115
+ "num_hidden_layers": 2,
116
+ "patch_size": 14,
117
+ "pixel_shuffle_ratio": 0.5,
118
+ "projector_dropout": 0.0,
119
+ "projector_input_dim": 32,
120
+ "projector_output_dim": 32,
121
+ "rope_theta": 10000,
122
+ "vision_feature_layer": -1,
123
+ "vision_feature_select_strategy": "default",
124
+ "vision_output_dim": 32
125
+ }
126
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/mixtral/dacorvo/Mixtral-tiny/80038e075c3313d11a1d.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
33
+ "on_device_sampling": false,
34
+ "optimum_neuron_version": "0.4.6.dev4",
35
+ "output_logits": false,
36
+ "pp_degree": 1,
37
+ "prefill_chunk_size": 0,
38
+ "sequence_length": 1024,
39
+ "speculation_length": 0,
40
+ "start_rank_id": 0,
41
+ "target": "trn1",
42
+ "torch_dtype": "float16",
43
+ "tp_degree": 2
44
+ },
45
+ "num_attention_heads": 32,
46
+ "num_experts_per_tok": 2,
47
+ "num_hidden_layers": 2,
48
+ "num_key_value_heads": 8,
49
+ "num_local_experts": 8,
50
+ "output_router_logits": false,
51
+ "rms_norm_eps": 1e-05,
52
+ "rope_theta": 10000.0,
53
+ "router_aux_loss_coef": 0.001,
54
+ "router_jitter_noise": 0.0,
55
+ "sliding_window": 4096,
56
+ "tie_word_embeddings": false,
57
+ "use_cache": true,
58
+ "vocab_size": 32000
59
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/phi3/yujiepan/phi-4-tiny-random/7d7f3f7850d2aa89e3fa.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.6.dev4",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "prefill_chunk_size": 0,
40
+ "sequence_length": 1024,
41
+ "speculation_length": 0,
42
+ "start_rank_id": 0,
43
+ "target": "trn1",
44
+ "torch_dtype": "bfloat16",
45
+ "tp_degree": 2
46
+ },
47
+ "num_attention_heads": 2,
48
+ "num_hidden_layers": 2,
49
+ "num_key_value_heads": 1,
50
+ "original_max_position_embeddings": 16384,
51
+ "partial_rotary_factor": 1.0,
52
+ "resid_pdrop": 0.0,
53
+ "rms_norm_eps": 1e-05,
54
+ "rope_scaling": null,
55
+ "rope_theta": 250000,
56
+ "sliding_window": null,
57
+ "tie_word_embeddings": false,
58
+ "use_cache": true,
59
+ "vocab_size": 100352
60
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/qwen2/Qwen/Qwen2.5-0.5B/e28a26007281cdef7599.json ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen2.5-0.5B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen2ForCausalLM"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "dtype": "bfloat16",
10
+ "hidden_act": "silu",
11
+ "hidden_size": 896,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 4864,
14
+ "layer_types": [
15
+ "full_attention",
16
+ "full_attention",
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
+ ],
40
+ "max_position_embeddings": 32768,
41
+ "max_window_layers": 24,
42
+ "model_type": "qwen2",
43
+ "neuron": {
44
+ "_serialized_key": "NxDNeuronConfig",
45
+ "batch_size": 2,
46
+ "capacity_factor": null,
47
+ "checkpoint_id": "Qwen/Qwen2.5-0.5B",
48
+ "checkpoint_revision": "060db6499f32faf8b98477b0a26969ef7d8b9987",
49
+ "continuous_batching": true,
50
+ "ep_degree": 1,
51
+ "fused_qkv": false,
52
+ "glu_mlp": true,
53
+ "local_ranks_size": 2,
54
+ "max_batch_size": 2,
55
+ "max_context_length": 1024,
56
+ "max_topk": 256,
57
+ "n_active_tokens": 1024,
58
+ "neuronxcc_version": "2.21.33363.0+82129205",
59
+ "on_device_sampling": false,
60
+ "optimum_neuron_version": "0.4.6.dev4",
61
+ "output_logits": false,
62
+ "pp_degree": 1,
63
+ "prefill_chunk_size": 0,
64
+ "sequence_length": 1024,
65
+ "speculation_length": 0,
66
+ "start_rank_id": 0,
67
+ "target": "trn1",
68
+ "torch_dtype": "bfloat16",
69
+ "tp_degree": 2
70
+ },
71
+ "num_attention_heads": 14,
72
+ "num_hidden_layers": 24,
73
+ "num_key_value_heads": 2,
74
+ "rms_norm_eps": 1e-06,
75
+ "rope_scaling": null,
76
+ "rope_theta": 1000000.0,
77
+ "sliding_window": null,
78
+ "tie_word_embeddings": true,
79
+ "use_cache": true,
80
+ "use_mrope": false,
81
+ "use_sliding_window": false,
82
+ "vocab_size": 151936
83
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/qwen2/yujiepan/qwen2.5-128k-tiny-random/c7e74a139ce290675aa6.json ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.6.dev4",
39
+ "output_logits": false,
40
+ "pp_degree": 1,
41
+ "prefill_chunk_size": 0,
42
+ "sequence_length": 1024,
43
+ "speculation_length": 0,
44
+ "start_rank_id": 0,
45
+ "target": "trn1",
46
+ "torch_dtype": "bfloat16",
47
+ "tp_degree": 2
48
+ },
49
+ "num_attention_heads": 4,
50
+ "num_hidden_layers": 2,
51
+ "num_key_value_heads": 2,
52
+ "rms_norm_eps": 1e-06,
53
+ "rope_scaling": {
54
+ "factor": 4.0,
55
+ "original_max_position_embeddings": 32768,
56
+ "rope_type": "yarn",
57
+ "type": "yarn"
58
+ },
59
+ "rope_theta": 1000000.0,
60
+ "sliding_window": null,
61
+ "tie_word_embeddings": false,
62
+ "use_cache": true,
63
+ "use_sliding_window": false,
64
+ "vocab_size": 152064
65
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/d6743412c8cf7830ae4a.json ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "2.21.33363.0+82129205",
38
+ "on_device_sampling": true,
39
+ "optimum_neuron_version": "0.4.6.dev4",
40
+ "output_logits": false,
41
+ "pp_degree": 1,
42
+ "prefill_chunk_size": 0,
43
+ "sequence_length": 1024,
44
+ "speculation_length": 0,
45
+ "start_rank_id": 0,
46
+ "target": "trn1",
47
+ "torch_dtype": "float32",
48
+ "tp_degree": 2
49
+ },
50
+ "norm_topk_prob": true,
51
+ "num_attention_heads": 2,
52
+ "num_experts": 8,
53
+ "num_experts_per_tok": 2,
54
+ "num_hidden_layers": 2,
55
+ "num_key_value_heads": 1,
56
+ "output_router_logits": false,
57
+ "rms_norm_eps": 1e-06,
58
+ "rope_scaling": null,
59
+ "rope_theta": 1000000.0,
60
+ "router_aux_loss_coef": 0.001,
61
+ "sliding_window": null,
62
+ "tie_word_embeddings": true,
63
+ "use_cache": true,
64
+ "use_sliding_window": false,
65
+ "vocab_size": 151936
66
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev4/smollm3/HuggingFaceTB/SmolLM3-3B/3aed09186384cc6540ef.json ADDED
@@ -0,0 +1,135 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "HuggingFaceTB/SmolLM3-3B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "SmolLM3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "hidden_act": "silu",
12
+ "hidden_size": 2048,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 11008,
15
+ "layer_types": [
16
+ "full_attention",
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
+ ],
53
+ "max_position_embeddings": 65536,
54
+ "max_window_layers": 28,
55
+ "mlp_bias": false,
56
+ "model_type": "smollm3",
57
+ "neuron": {
58
+ "_serialized_key": "NxDNeuronConfig",
59
+ "batch_size": 1,
60
+ "capacity_factor": null,
61
+ "checkpoint_id": "HuggingFaceTB/SmolLM3-3B",
62
+ "checkpoint_revision": "a07cc9a04f16550a088caea529712d1d335b0ac1",
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": 1024,
70
+ "max_topk": 256,
71
+ "n_active_tokens": 1024,
72
+ "neuronxcc_version": "2.21.33363.0+82129205",
73
+ "on_device_sampling": true,
74
+ "optimum_neuron_version": "0.4.6.dev4",
75
+ "output_logits": false,
76
+ "pp_degree": 1,
77
+ "prefill_chunk_size": 0,
78
+ "sequence_length": 1024,
79
+ "speculation_length": 0,
80
+ "start_rank_id": 0,
81
+ "target": "trn1",
82
+ "torch_dtype": "bfloat16",
83
+ "tp_degree": 2
84
+ },
85
+ "no_rope_layer_interval": 4,
86
+ "no_rope_layers": [
87
+ 1,
88
+ 1,
89
+ 1,
90
+ 0,
91
+ 1,
92
+ 1,
93
+ 1,
94
+ 0,
95
+ 1,
96
+ 1,
97
+ 1,
98
+ 0,
99
+ 1,
100
+ 1,
101
+ 1,
102
+ 0,
103
+ 1,
104
+ 1,
105
+ 1,
106
+ 0,
107
+ 1,
108
+ 1,
109
+ 1,
110
+ 0,
111
+ 1,
112
+ 1,
113
+ 1,
114
+ 0,
115
+ 1,
116
+ 1,
117
+ 1,
118
+ 0,
119
+ 1,
120
+ 1,
121
+ 1,
122
+ 0
123
+ ],
124
+ "num_attention_heads": 16,
125
+ "num_hidden_layers": 36,
126
+ "num_key_value_heads": 4,
127
+ "pretraining_tp": 2,
128
+ "rms_norm_eps": 1e-06,
129
+ "rope_scaling": null,
130
+ "rope_theta": 5000000.0,
131
+ "sliding_window": null,
132
+ "use_cache": false,
133
+ "use_sliding_window": false,
134
+ "vocab_size": 128256
135
+ }
neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/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_vision_encoder/vision_encoder/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b868910b39583cee3aeeab326d9fb6bfc1e12ce4e937d244abc110c48b64465e
3
+ size 160725
neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ae1f3f7cb2cc80137f4849385dad6bb0743f033cf7ede6d04c9dc4081d8e1d82
3
+ size 30598144
neuronxcc-2.21.33363.0+82129205/MODULE_04c64527506049d877d9+b02446f6/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e4fc9bb782609dfaeb2b74d4fdc0108e917006c4f1622b3a05d2f0ebefc20b16
3
+ size 30708309
neuronxcc-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+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-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+a02c3a36/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+a02c3a36/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2be0f520e77dfad494fc89a6f2bda248a6149afe1faaa16ff57adc5899bcefb2
3
+ size 686960
neuronxcc-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+a02c3a36/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:422a22826c569c92260c9e11e4c7845725eb13de7f0943b0e6f012f3908c996b
3
+ size 1485824
neuronxcc-2.21.33363.0+82129205/MODULE_789b4718541dd2474859+a02c3a36/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f412c2af4dc5c77c58211597836cecb60ff04a8741197af646a6c6b39fc137ab
3
+ size 1596702
neuronxcc-2.21.33363.0+82129205/MODULE_82382ce0b34b3e272481+6170d8e1/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/chunked_prefill/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.33363.0+82129205/MODULE_82382ce0b34b3e272481+6170d8e1/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_82382ce0b34b3e272481+6170d8e1/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:98739a3fd726ac3a8379af0d0b6e611702ebd66977e60a994ce9155f14e9cb3e
3
+ size 911470
neuronxcc-2.21.33363.0+82129205/MODULE_82382ce0b34b3e272481+6170d8e1/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68e1150370b99c7ff1054699ec25421e64872fd7e32e9777e95cefc3c4e8e4a2
3
+ size 11674624