Hugging Face's logo

optimum-internal-testing
/
neuron-testing-cache

Model card Files
xet
 Community
tengomucho HF Staff commited on
Commit
b35f1cb
·
verified ·
1 Parent(s): 2b84f66

Synchronizing local compiler cache.

Browse files
Files changed (43) hide show
  1. .gitattributes +3 -0
  2. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/03b0c107d1cede36875199a5d51decfe04c473de2af9999f8577a028d74d0ab4/566b663bde0d89e24b29.json +188 -0
  3. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/2da7a00f0478d50ae1e7f75f085c5b2773b5f355f427c61cf34cb6febd629d96/e5ddd3afb102c02baf93.json +59 -0
  4. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/441269935591cad8d370e512c0b93cdd2fce6247c40e5a4866d872ee5338b0de/51eb5f08405da966cefd.json +81 -0
  5. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/4ab8140bc7eb4a553d95855c5c2be2cf8c0fbab21b823d76183b6f51e98b6fc5/03a64a22d1b885eece61.json +58 -0
  6. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/6454afdf3e9d66c7226c13a575b718845c25e53b0699600ba2bb4f883e9d841b/97e03fcfc7f46ac8836e.json +62 -0
  7. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/73707b485eab9008c7aba7f5dad0ce2384ac685318d5f888c12fa0d81ed90b19/d04c2a3f18746af5f901.json +134 -0
  8. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/7518518c7e077820070186deda960d8cc49db068cdf0ac70664098fa2b6b698c/29c61ad2f54baaec4c1d.json +64 -0
  9. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/7f05bde17c7b0ffeb657897697f23d182f406b76ced7f1b2cd5741dc93fe2e2e/dadf01a9f544218eabdd.json +125 -0
  10. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/920f44ce6d3e004d1ce547ae06644f7be262180644b04573153aa15d98742edc/b5678f2b1f926f36a4fd.json +65 -0
  11. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/d139acf64685f15794bb983ff6eb881bdd31304bae88b0ce1ed20a54c21f2265/0e7a6f2933f99785cba6.json +58 -0
  12. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/0e7a6f2933f99785cba6.json +58 -0
  13. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/idefics3/HuggingFaceTB/SmolVLM-256M-Instruct/566b663bde0d89e24b29.json +188 -0
  14. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/llama/llamafactory/tiny-random-Llama-3/97e03fcfc7f46ac8836e.json +62 -0
  15. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/llama4/tiny-random/llama-4/dadf01a9f544218eabdd.json +125 -0
  16. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/mixtral/dacorvo/Mixtral-tiny/03a64a22d1b885eece61.json +58 -0
  17. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/phi3/yujiepan/phi-4-tiny-random/e5ddd3afb102c02baf93.json +59 -0
  18. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/qwen2/yujiepan/qwen2.5-128k-tiny-random/29c61ad2f54baaec4c1d.json +64 -0
  19. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/b5678f2b1f926f36a4fd.json +65 -0
  20. neuronxcc-2.21.33363.0+82129205/MODULE_17476209562158013765+e30acd3a/model.neff +0 -0
  21. neuronxcc-2.21.33363.0+82129205/MODULE_18b928d4e4846c6ce9c7+24129607/compile_flags.json +1 -0
  22. neuronxcc-2.21.33363.0+82129205/MODULE_18b928d4e4846c6ce9c7+24129607/model.done +0 -0
  23. neuronxcc-2.21.33363.0+82129205/MODULE_18b928d4e4846c6ce9c7+24129607/model.hlo_module.pb +3 -0
  24. neuronxcc-2.21.33363.0+82129205/MODULE_18b928d4e4846c6ce9c7+24129607/model.neff +3 -0
  25. neuronxcc-2.21.33363.0+82129205/MODULE_27b2c4720a6a1662f8b5+24129607/model.hlo_module.pb +1 -1
  26. neuronxcc-2.21.33363.0+82129205/MODULE_27b2c4720a6a1662f8b5+24129607/model.neff +1 -1
  27. neuronxcc-2.21.33363.0+82129205/MODULE_50933980a498d1588ed1+a02c3a36/compile_flags.json +1 -0
  28. neuronxcc-2.21.33363.0+82129205/MODULE_50933980a498d1588ed1+a02c3a36/model.done +0 -0
  29. neuronxcc-2.21.33363.0+82129205/MODULE_50933980a498d1588ed1+a02c3a36/model.hlo_module.pb +3 -0
  30. neuronxcc-2.21.33363.0+82129205/MODULE_50933980a498d1588ed1+a02c3a36/model.neff +3 -0
  31. neuronxcc-2.21.33363.0+82129205/MODULE_50933980a498d1588ed1+a02c3a36/wrapped_neff.hlo +3 -0
  32. neuronxcc-2.21.33363.0+82129205/MODULE_60f37ef2cb181ebeb434+a02c3a36/model.neff +1 -1
  33. neuronxcc-2.21.33363.0+82129205/MODULE_60f37ef2cb181ebeb434+a02c3a36/wrapped_neff.hlo +1 -1
  34. neuronxcc-2.21.33363.0+82129205/MODULE_95b9072f246645f24461+a02c3a36/model.hlo_module.pb +1 -1
  35. neuronxcc-2.21.33363.0+82129205/MODULE_95b9072f246645f24461+a02c3a36/model.neff +1 -1
  36. neuronxcc-2.21.33363.0+82129205/MODULE_95b9072f246645f24461+a02c3a36/wrapped_neff.hlo +1 -1
  37. neuronxcc-2.21.33363.0+82129205/MODULE_a1cd38756c74f5816b8b+24129607/model.hlo_module.pb +1 -1
  38. neuronxcc-2.21.33363.0+82129205/MODULE_a1cd38756c74f5816b8b+24129607/model.neff +1 -1
  39. neuronxcc-2.21.33363.0+82129205/MODULE_ad16674dd3fc5a51c226+a02c3a36/model.neff +1 -1
  40. neuronxcc-2.21.33363.0+82129205/MODULE_ad16674dd3fc5a51c226+a02c3a36/wrapped_neff.hlo +1 -1
  41. neuronxcc-2.21.33363.0+82129205/MODULE_c69743a3b026cb8c0f9e+a02c3a36/model.hlo_module.pb +1 -1
  42. neuronxcc-2.21.33363.0+82129205/MODULE_c69743a3b026cb8c0f9e+a02c3a36/model.neff +1 -1
  43. neuronxcc-2.21.33363.0+82129205/MODULE_c69743a3b026cb8c0f9e+a02c3a36/wrapped_neff.hlo +1 -1
.gitattributes CHANGED
@@ -7135,3 +7135,6 @@ neuronxcc-2.21.33363.0+82129205/MODULE_d50f09930f3886ac4b66+a02c3a36/wrapped_nef
7135
  neuronxcc-2.21.33363.0+82129205/MODULE_1d8d646ed52cd9b3ce0f+6170d8e1/model.neff filter=lfs diff=lfs merge=lfs -text
7136
  neuronxcc-2.21.33363.0+82129205/MODULE_71833bfb10ff3967bd3d+6170d8e1/model.neff filter=lfs diff=lfs merge=lfs -text
7137
  neuronxcc-2.21.33363.0+82129205/MODULE_da7543b73fbd9c77b695+6170d8e1/model.neff filter=lfs diff=lfs merge=lfs -text
 
 
 
 
7135
  neuronxcc-2.21.33363.0+82129205/MODULE_1d8d646ed52cd9b3ce0f+6170d8e1/model.neff filter=lfs diff=lfs merge=lfs -text
7136
  neuronxcc-2.21.33363.0+82129205/MODULE_71833bfb10ff3967bd3d+6170d8e1/model.neff filter=lfs diff=lfs merge=lfs -text
7137
  neuronxcc-2.21.33363.0+82129205/MODULE_da7543b73fbd9c77b695+6170d8e1/model.neff filter=lfs diff=lfs merge=lfs -text
7138
+ neuronxcc-2.21.33363.0+82129205/MODULE_18b928d4e4846c6ce9c7+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
7139
+ neuronxcc-2.21.33363.0+82129205/MODULE_50933980a498d1588ed1+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
7140
+ neuronxcc-2.21.33363.0+82129205/MODULE_50933980a498d1588ed1+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/03b0c107d1cede36875199a5d51decfe04c473de2af9999f8577a028d74d0ab4/566b663bde0d89e24b29.json ADDED
@@ -0,0 +1,188 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 1,
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.dev3",
32
+ "output_logits": false,
33
+ "pp_degree": 1,
34
+ "sequence_length": 2048,
35
+ "speculation_length": 0,
36
+ "start_rank_id": 0,
37
+ "target": "trn1",
38
+ "torch_dtype": "bfloat16",
39
+ "tp_degree": 2
40
+ },
41
+ "scale_factor": 4,
42
+ "text_config": {
43
+ "_attn_implementation_autoset": false,
44
+ "_flash_attn_2_enabled": true,
45
+ "_name_or_path": "None",
46
+ "architectures": [
47
+ "VLlama3ForCausalLM"
48
+ ],
49
+ "attention_bias": false,
50
+ "attention_dropout": 0.0,
51
+ "dtype": "bfloat16",
52
+ "head_dim": 64,
53
+ "hidden_act": "silu",
54
+ "hidden_size": 576,
55
+ "initializer_range": 0.041666666666666664,
56
+ "intermediate_size": 1536,
57
+ "is_llama_config": true,
58
+ "max_position_embeddings": 8192,
59
+ "mlp_bias": false,
60
+ "model_type": "llama",
61
+ "neftune_noise_alpha": 0.0,
62
+ "num_attention_heads": 9,
63
+ "num_hidden_layers": 30,
64
+ "num_key_value_heads": 3,
65
+ "pad_token_id": 2,
66
+ "perceiver_config": {
67
+ "_attn_implementation_autoset": false,
68
+ "_name_or_path": "",
69
+ "add_cross_attention": false,
70
+ "architectures": null,
71
+ "attention_dropout": 0.0,
72
+ "bad_words_ids": null,
73
+ "begin_suppress_tokens": null,
74
+ "bos_token_id": null,
75
+ "chunk_size_feed_forward": 0,
76
+ "cross_attention_hidden_size": null,
77
+ "decoder_start_token_id": null,
78
+ "diversity_penalty": 0.0,
79
+ "do_sample": false,
80
+ "early_stopping": false,
81
+ "encoder_no_repeat_ngram_size": 0,
82
+ "eos_token_id": null,
83
+ "exponential_decay_length_penalty": null,
84
+ "finetuning_task": null,
85
+ "forced_bos_token_id": null,
86
+ "forced_eos_token_id": null,
87
+ "hidden_act": "silu",
88
+ "id2label": {
89
+ "0": "LABEL_0",
90
+ "1": "LABEL_1"
91
+ },
92
+ "is_decoder": false,
93
+ "is_encoder_decoder": false,
94
+ "label2id": {
95
+ "LABEL_0": 0,
96
+ "LABEL_1": 1
97
+ },
98
+ "length_penalty": 1.0,
99
+ "max_length": 20,
100
+ "min_length": 0,
101
+ "model_type": "vllama3",
102
+ "no_repeat_ngram_size": 0,
103
+ "num_beam_groups": 1,
104
+ "num_beams": 1,
105
+ "num_key_value_heads": 1,
106
+ "num_return_sequences": 1,
107
+ "output_attentions": false,
108
+ "output_hidden_states": false,
109
+ "output_scores": false,
110
+ "pad_token_id": null,
111
+ "prefix": null,
112
+ "problem_type": null,
113
+ "pruned_heads": {},
114
+ "qk_layer_norms_perceiver": false,
115
+ "remove_invalid_values": false,
116
+ "repetition_penalty": 1.0,
117
+ "resampler_depth": 6,
118
+ "resampler_head_dim": 96,
119
+ "resampler_n_heads": 16,
120
+ "resampler_n_latents": 64,
121
+ "return_dict": true,
122
+ "return_dict_in_generate": false,
123
+ "sep_token_id": null,
124
+ "suppress_tokens": null,
125
+ "task_specific_params": null,
126
+ "temperature": 1.0,
127
+ "tf_legacy_loss": false,
128
+ "tie_encoder_decoder": false,
129
+ "tie_word_embeddings": true,
130
+ "tokenizer_class": null,
131
+ "top_k": 50,
132
+ "top_p": 1.0,
133
+ "torch_dtype": null,
134
+ "torchscript": false,
135
+ "transformers_version": "4.46.0",
136
+ "typical_p": 1.0,
137
+ "use_bfloat16": false
138
+ },
139
+ "pixel_shuffle_factor": 4,
140
+ "pretraining_tp": 1,
141
+ "qk_layer_norms": false,
142
+ "rms_norm_eps": 1e-05,
143
+ "rope_interleaved": false,
144
+ "rope_scaling": null,
145
+ "rope_theta": 100000,
146
+ "transformers.js_config": {
147
+ "kv_cache_dtype": {
148
+ "fp16": "float16",
149
+ "q4f16": "float16"
150
+ }
151
+ },
152
+ "use_cache": true,
153
+ "use_resampler": false,
154
+ "vocab_size": 49280
155
+ },
156
+ "tie_word_embeddings": false,
157
+ "transformers.js_config": {
158
+ "kv_cache_dtype": {
159
+ "fp16": "float16",
160
+ "q4f16": "float16"
161
+ }
162
+ },
163
+ "use_cache": true,
164
+ "vision_config": {
165
+ "_attn_implementation_autoset": false,
166
+ "attention_dropout": 0.0,
167
+ "hidden_act": "gelu_pytorch_tanh",
168
+ "hidden_size": 768,
169
+ "image_size": 512,
170
+ "initializer_range": 0.02,
171
+ "intermediate_size": 3072,
172
+ "layer_norm_eps": 1e-06,
173
+ "max_image_size": {
174
+ "longest_edge": 512
175
+ },
176
+ "model_type": "idefics3_vision",
177
+ "num_attention_heads": 12,
178
+ "num_channels": 3,
179
+ "num_hidden_layers": 12,
180
+ "patch_size": 16,
181
+ "size": {
182
+ "longest_edge": 2048
183
+ },
184
+ "tie_word_embeddings": false,
185
+ "use_base_siglip": true
186
+ },
187
+ "vocab_size": 49280
188
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/2da7a00f0478d50ae1e7f75f085c5b2773b5f355f427c61cf34cb6febd629d96/e5ddd3afb102c02baf93.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": "2.21.33363.0+82129205",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/441269935591cad8d370e512c0b93cdd2fce6247c40e5a4866d872ee5338b0de/51eb5f08405da966cefd.json ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.dev3",
59
+ "output_logits": false,
60
+ "pp_degree": 1,
61
+ "sequence_length": 1024,
62
+ "speculation_length": 0,
63
+ "start_rank_id": 0,
64
+ "target": "trn1",
65
+ "torch_dtype": "bfloat16",
66
+ "tp_degree": 2
67
+ },
68
+ "num_attention_heads": 4,
69
+ "num_hidden_layers": 18,
70
+ "num_key_value_heads": 1,
71
+ "query_pre_attn_scalar": 256,
72
+ "rms_norm_eps": 1e-06,
73
+ "rope_local_base_freq": 10000.0,
74
+ "rope_scaling": null,
75
+ "rope_theta": 1000000.0,
76
+ "sliding_window": 512,
77
+ "unsloth_fixed": true,
78
+ "use_bidirectional_attention": false,
79
+ "use_cache": true,
80
+ "vocab_size": 262144
81
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/4ab8140bc7eb4a553d95855c5c2be2cf8c0fbab21b823d76183b6f51e98b6fc5/03a64a22d1b885eece61.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": "2.21.33363.0+82129205",
33
+ "on_device_sampling": false,
34
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/6454afdf3e9d66c7226c13a575b718845c25e53b0699600ba2bb4f883e9d841b/97e03fcfc7f46ac8836e.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": "2.21.33363.0+82129205",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/73707b485eab9008c7aba7f5dad0ce2384ac685318d5f888c12fa0d81ed90b19/d04c2a3f18746af5f901.json ADDED
@@ -0,0 +1,134 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.dev3",
75
+ "output_logits": false,
76
+ "pp_degree": 1,
77
+ "sequence_length": 1024,
78
+ "speculation_length": 0,
79
+ "start_rank_id": 0,
80
+ "target": "trn1",
81
+ "torch_dtype": "bfloat16",
82
+ "tp_degree": 2
83
+ },
84
+ "no_rope_layer_interval": 4,
85
+ "no_rope_layers": [
86
+ 1,
87
+ 1,
88
+ 1,
89
+ 0,
90
+ 1,
91
+ 1,
92
+ 1,
93
+ 0,
94
+ 1,
95
+ 1,
96
+ 1,
97
+ 0,
98
+ 1,
99
+ 1,
100
+ 1,
101
+ 0,
102
+ 1,
103
+ 1,
104
+ 1,
105
+ 0,
106
+ 1,
107
+ 1,
108
+ 1,
109
+ 0,
110
+ 1,
111
+ 1,
112
+ 1,
113
+ 0,
114
+ 1,
115
+ 1,
116
+ 1,
117
+ 0,
118
+ 1,
119
+ 1,
120
+ 1,
121
+ 0
122
+ ],
123
+ "num_attention_heads": 16,
124
+ "num_hidden_layers": 36,
125
+ "num_key_value_heads": 4,
126
+ "pretraining_tp": 2,
127
+ "rms_norm_eps": 1e-06,
128
+ "rope_scaling": null,
129
+ "rope_theta": 5000000.0,
130
+ "sliding_window": null,
131
+ "use_cache": false,
132
+ "use_sliding_window": false,
133
+ "vocab_size": 128256
134
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/7518518c7e077820070186deda960d8cc49db068cdf0ac70664098fa2b6b698c/29c61ad2f54baaec4c1d.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": "2.21.33363.0+82129205",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/7f05bde17c7b0ffeb657897697f23d182f406b76ced7f1b2cd5741dc93fe2e2e/dadf01a9f544218eabdd.json ADDED
@@ -0,0 +1,125 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.dev3",
31
+ "output_logits": false,
32
+ "pp_degree": 1,
33
+ "sequence_length": 1024,
34
+ "speculation_length": 0,
35
+ "start_rank_id": 0,
36
+ "target": "trn1",
37
+ "torch_dtype": "bfloat16",
38
+ "tp_degree": 2
39
+ },
40
+ "text_config": {
41
+ "_attn_implementation_autoset": true,
42
+ "attention_bias": false,
43
+ "attention_chunk_size": 128,
44
+ "attention_dropout": 0.0,
45
+ "attn_scale": 0.1,
46
+ "attn_temperature_tuning": 4,
47
+ "bos_token_id": 200000,
48
+ "cache_implementation": "hybrid",
49
+ "dtype": "bfloat16",
50
+ "eos_token_id": [
51
+ 200001,
52
+ 200007,
53
+ 200008
54
+ ],
55
+ "floor_scale": 8192,
56
+ "for_llm_compressor": false,
57
+ "head_dim": 32,
58
+ "hidden_act": "silu",
59
+ "hidden_size": 32,
60
+ "initializer_range": 0.02,
61
+ "interleave_moe_layer_step": 2,
62
+ "intermediate_size": 64,
63
+ "intermediate_size_mlp": 128,
64
+ "layer_types": [
65
+ "chunked_attention",
66
+ "chunked_attention",
67
+ "chunked_attention",
68
+ "full_attention"
69
+ ],
70
+ "max_position_embeddings": 1048576,
71
+ "model_type": "llama4_text",
72
+ "moe_layers": [
73
+ 1,
74
+ 3
75
+ ],
76
+ "no_rope_layers": [
77
+ 1,
78
+ 1,
79
+ 1,
80
+ 0
81
+ ],
82
+ "num_attention_heads": 1,
83
+ "num_experts_per_tok": 1,
84
+ "num_hidden_layers": 4,
85
+ "num_key_value_heads": 1,
86
+ "num_local_experts": 8,
87
+ "output_router_logits": false,
88
+ "pad_token_id": 200018,
89
+ "rms_norm_eps": 1e-05,
90
+ "rope_scaling": null,
91
+ "rope_theta": 500000.0,
92
+ "router_aux_loss_coef": 0.001,
93
+ "router_jitter_noise": 0.0,
94
+ "tie_word_embeddings": true,
95
+ "use_cache": true,
96
+ "use_qk_norm": true,
97
+ "vocab_size": 202048
98
+ },
99
+ "tie_word_embeddings": false,
100
+ "vision_config": {
101
+ "_attn_implementation_autoset": true,
102
+ "_vision_feature_layer": -1,
103
+ "attention_dropout": 0.0,
104
+ "hidden_act": "gelu",
105
+ "hidden_size": 32,
106
+ "image_size": 336,
107
+ "initializer_range": 0.02,
108
+ "intermediate_size": 128,
109
+ "model_type": "llama4_vision_model",
110
+ "multi_modal_projector_bias": false,
111
+ "norm_eps": 1e-05,
112
+ "num_attention_heads": 1,
113
+ "num_channels": 3,
114
+ "num_hidden_layers": 2,
115
+ "patch_size": 14,
116
+ "pixel_shuffle_ratio": 0.5,
117
+ "projector_dropout": 0.0,
118
+ "projector_input_dim": 32,
119
+ "projector_output_dim": 32,
120
+ "rope_theta": 10000,
121
+ "vision_feature_layer": -1,
122
+ "vision_feature_select_strategy": "default",
123
+ "vision_output_dim": 32
124
+ }
125
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/920f44ce6d3e004d1ce547ae06644f7be262180644b04573153aa15d98742edc/b5678f2b1f926f36a4fd.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": "2.21.33363.0+82129205",
38
+ "on_device_sampling": true,
39
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/d139acf64685f15794bb983ff6eb881bdd31304bae88b0ce1ed20a54c21f2265/0e7a6f2933f99785cba6.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": "2.21.33363.0+82129205",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/0e7a6f2933f99785cba6.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": "2.21.33363.0+82129205",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/idefics3/HuggingFaceTB/SmolVLM-256M-Instruct/566b663bde0d89e24b29.json ADDED
@@ -0,0 +1,188 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 1,
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.dev3",
32
+ "output_logits": false,
33
+ "pp_degree": 1,
34
+ "sequence_length": 2048,
35
+ "speculation_length": 0,
36
+ "start_rank_id": 0,
37
+ "target": "trn1",
38
+ "torch_dtype": "bfloat16",
39
+ "tp_degree": 2
40
+ },
41
+ "scale_factor": 4,
42
+ "text_config": {
43
+ "_attn_implementation_autoset": false,
44
+ "_flash_attn_2_enabled": true,
45
+ "_name_or_path": "None",
46
+ "architectures": [
47
+ "VLlama3ForCausalLM"
48
+ ],
49
+ "attention_bias": false,
50
+ "attention_dropout": 0.0,
51
+ "dtype": "bfloat16",
52
+ "head_dim": 64,
53
+ "hidden_act": "silu",
54
+ "hidden_size": 576,
55
+ "initializer_range": 0.041666666666666664,
56
+ "intermediate_size": 1536,
57
+ "is_llama_config": true,
58
+ "max_position_embeddings": 8192,
59
+ "mlp_bias": false,
60
+ "model_type": "llama",
61
+ "neftune_noise_alpha": 0.0,
62
+ "num_attention_heads": 9,
63
+ "num_hidden_layers": 30,
64
+ "num_key_value_heads": 3,
65
+ "pad_token_id": 2,
66
+ "perceiver_config": {
67
+ "_attn_implementation_autoset": false,
68
+ "_name_or_path": "",
69
+ "add_cross_attention": false,
70
+ "architectures": null,
71
+ "attention_dropout": 0.0,
72
+ "bad_words_ids": null,
73
+ "begin_suppress_tokens": null,
74
+ "bos_token_id": null,
75
+ "chunk_size_feed_forward": 0,
76
+ "cross_attention_hidden_size": null,
77
+ "decoder_start_token_id": null,
78
+ "diversity_penalty": 0.0,
79
+ "do_sample": false,
80
+ "early_stopping": false,
81
+ "encoder_no_repeat_ngram_size": 0,
82
+ "eos_token_id": null,
83
+ "exponential_decay_length_penalty": null,
84
+ "finetuning_task": null,
85
+ "forced_bos_token_id": null,
86
+ "forced_eos_token_id": null,
87
+ "hidden_act": "silu",
88
+ "id2label": {
89
+ "0": "LABEL_0",
90
+ "1": "LABEL_1"
91
+ },
92
+ "is_decoder": false,
93
+ "is_encoder_decoder": false,
94
+ "label2id": {
95
+ "LABEL_0": 0,
96
+ "LABEL_1": 1
97
+ },
98
+ "length_penalty": 1.0,
99
+ "max_length": 20,
100
+ "min_length": 0,
101
+ "model_type": "vllama3",
102
+ "no_repeat_ngram_size": 0,
103
+ "num_beam_groups": 1,
104
+ "num_beams": 1,
105
+ "num_key_value_heads": 1,
106
+ "num_return_sequences": 1,
107
+ "output_attentions": false,
108
+ "output_hidden_states": false,
109
+ "output_scores": false,
110
+ "pad_token_id": null,
111
+ "prefix": null,
112
+ "problem_type": null,
113
+ "pruned_heads": {},
114
+ "qk_layer_norms_perceiver": false,
115
+ "remove_invalid_values": false,
116
+ "repetition_penalty": 1.0,
117
+ "resampler_depth": 6,
118
+ "resampler_head_dim": 96,
119
+ "resampler_n_heads": 16,
120
+ "resampler_n_latents": 64,
121
+ "return_dict": true,
122
+ "return_dict_in_generate": false,
123
+ "sep_token_id": null,
124
+ "suppress_tokens": null,
125
+ "task_specific_params": null,
126
+ "temperature": 1.0,
127
+ "tf_legacy_loss": false,
128
+ "tie_encoder_decoder": false,
129
+ "tie_word_embeddings": true,
130
+ "tokenizer_class": null,
131
+ "top_k": 50,
132
+ "top_p": 1.0,
133
+ "torch_dtype": null,
134
+ "torchscript": false,
135
+ "transformers_version": "4.46.0",
136
+ "typical_p": 1.0,
137
+ "use_bfloat16": false
138
+ },
139
+ "pixel_shuffle_factor": 4,
140
+ "pretraining_tp": 1,
141
+ "qk_layer_norms": false,
142
+ "rms_norm_eps": 1e-05,
143
+ "rope_interleaved": false,
144
+ "rope_scaling": null,
145
+ "rope_theta": 100000,
146
+ "transformers.js_config": {
147
+ "kv_cache_dtype": {
148
+ "fp16": "float16",
149
+ "q4f16": "float16"
150
+ }
151
+ },
152
+ "use_cache": true,
153
+ "use_resampler": false,
154
+ "vocab_size": 49280
155
+ },
156
+ "tie_word_embeddings": false,
157
+ "transformers.js_config": {
158
+ "kv_cache_dtype": {
159
+ "fp16": "float16",
160
+ "q4f16": "float16"
161
+ }
162
+ },
163
+ "use_cache": true,
164
+ "vision_config": {
165
+ "_attn_implementation_autoset": false,
166
+ "attention_dropout": 0.0,
167
+ "hidden_act": "gelu_pytorch_tanh",
168
+ "hidden_size": 768,
169
+ "image_size": 512,
170
+ "initializer_range": 0.02,
171
+ "intermediate_size": 3072,
172
+ "layer_norm_eps": 1e-06,
173
+ "max_image_size": {
174
+ "longest_edge": 512
175
+ },
176
+ "model_type": "idefics3_vision",
177
+ "num_attention_heads": 12,
178
+ "num_channels": 3,
179
+ "num_hidden_layers": 12,
180
+ "patch_size": 16,
181
+ "size": {
182
+ "longest_edge": 2048
183
+ },
184
+ "tie_word_embeddings": false,
185
+ "use_base_siglip": true
186
+ },
187
+ "vocab_size": 49280
188
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/llama/llamafactory/tiny-random-Llama-3/97e03fcfc7f46ac8836e.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": "2.21.33363.0+82129205",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/llama4/tiny-random/llama-4/dadf01a9f544218eabdd.json ADDED
@@ -0,0 +1,125 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.dev3",
31
+ "output_logits": false,
32
+ "pp_degree": 1,
33
+ "sequence_length": 1024,
34
+ "speculation_length": 0,
35
+ "start_rank_id": 0,
36
+ "target": "trn1",
37
+ "torch_dtype": "bfloat16",
38
+ "tp_degree": 2
39
+ },
40
+ "text_config": {
41
+ "_attn_implementation_autoset": true,
42
+ "attention_bias": false,
43
+ "attention_chunk_size": 128,
44
+ "attention_dropout": 0.0,
45
+ "attn_scale": 0.1,
46
+ "attn_temperature_tuning": 4,
47
+ "bos_token_id": 200000,
48
+ "cache_implementation": "hybrid",
49
+ "dtype": "bfloat16",
50
+ "eos_token_id": [
51
+ 200001,
52
+ 200007,
53
+ 200008
54
+ ],
55
+ "floor_scale": 8192,
56
+ "for_llm_compressor": false,
57
+ "head_dim": 32,
58
+ "hidden_act": "silu",
59
+ "hidden_size": 32,
60
+ "initializer_range": 0.02,
61
+ "interleave_moe_layer_step": 2,
62
+ "intermediate_size": 64,
63
+ "intermediate_size_mlp": 128,
64
+ "layer_types": [
65
+ "chunked_attention",
66
+ "chunked_attention",
67
+ "chunked_attention",
68
+ "full_attention"
69
+ ],
70
+ "max_position_embeddings": 1048576,
71
+ "model_type": "llama4_text",
72
+ "moe_layers": [
73
+ 1,
74
+ 3
75
+ ],
76
+ "no_rope_layers": [
77
+ 1,
78
+ 1,
79
+ 1,
80
+ 0
81
+ ],
82
+ "num_attention_heads": 1,
83
+ "num_experts_per_tok": 1,
84
+ "num_hidden_layers": 4,
85
+ "num_key_value_heads": 1,
86
+ "num_local_experts": 8,
87
+ "output_router_logits": false,
88
+ "pad_token_id": 200018,
89
+ "rms_norm_eps": 1e-05,
90
+ "rope_scaling": null,
91
+ "rope_theta": 500000.0,
92
+ "router_aux_loss_coef": 0.001,
93
+ "router_jitter_noise": 0.0,
94
+ "tie_word_embeddings": true,
95
+ "use_cache": true,
96
+ "use_qk_norm": true,
97
+ "vocab_size": 202048
98
+ },
99
+ "tie_word_embeddings": false,
100
+ "vision_config": {
101
+ "_attn_implementation_autoset": true,
102
+ "_vision_feature_layer": -1,
103
+ "attention_dropout": 0.0,
104
+ "hidden_act": "gelu",
105
+ "hidden_size": 32,
106
+ "image_size": 336,
107
+ "initializer_range": 0.02,
108
+ "intermediate_size": 128,
109
+ "model_type": "llama4_vision_model",
110
+ "multi_modal_projector_bias": false,
111
+ "norm_eps": 1e-05,
112
+ "num_attention_heads": 1,
113
+ "num_channels": 3,
114
+ "num_hidden_layers": 2,
115
+ "patch_size": 14,
116
+ "pixel_shuffle_ratio": 0.5,
117
+ "projector_dropout": 0.0,
118
+ "projector_input_dim": 32,
119
+ "projector_output_dim": 32,
120
+ "rope_theta": 10000,
121
+ "vision_feature_layer": -1,
122
+ "vision_feature_select_strategy": "default",
123
+ "vision_output_dim": 32
124
+ }
125
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/mixtral/dacorvo/Mixtral-tiny/03a64a22d1b885eece61.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": "2.21.33363.0+82129205",
33
+ "on_device_sampling": false,
34
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/phi3/yujiepan/phi-4-tiny-random/e5ddd3afb102c02baf93.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": "2.21.33363.0+82129205",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/qwen2/yujiepan/qwen2.5-128k-tiny-random/29c61ad2f54baaec4c1d.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": "2.21.33363.0+82129205",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev3/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/b5678f2b1f926f36a4fd.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": "2.21.33363.0+82129205",
38
+ "on_device_sampling": true,
39
+ "optimum_neuron_version": "0.4.6.dev3",
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-2.21.33363.0+82129205/MODULE_17476209562158013765+e30acd3a/model.neff CHANGED
Binary files a/neuronxcc-2.21.33363.0+82129205/MODULE_17476209562158013765+e30acd3a/model.neff and b/neuronxcc-2.21.33363.0+82129205/MODULE_17476209562158013765+e30acd3a/model.neff differ
 
neuronxcc-2.21.33363.0+82129205/MODULE_18b928d4e4846c6ce9c7+24129607/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/context_encoding/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.33363.0+82129205/MODULE_18b928d4e4846c6ce9c7+24129607/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_18b928d4e4846c6ce9c7+24129607/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d06e13fb7c050bb1840d5521e950d28037d97bc6b245f8ac9dd44e23e1cbfeb8
3
+ size 597952
neuronxcc-2.21.33363.0+82129205/MODULE_18b928d4e4846c6ce9c7+24129607/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c4da2c1a0a0177b946afd719d4b1361c4458c97c02a2dd404f8e7cff8aa5abc9
3
+ size 7394304
neuronxcc-2.21.33363.0+82129205/MODULE_27b2c4720a6a1662f8b5+24129607/model.hlo_module.pb CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1dddf91c6a86546b55890eda4dd0026330a7ea59d853f39f8ac2758b366729a6
3
  size 961456
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc270384625da985be79e1806b1b95e3c06eaccdcbfc110c85006b28c1f920f9
3
  size 961456
neuronxcc-2.21.33363.0+82129205/MODULE_27b2c4720a6a1662f8b5+24129607/model.neff CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9c9b303d5ffb9dbe60b8a13f4198948947628044e3564750dd748a1831dcfceb
3
  size 14654464
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6dfd1091bff77d12e9c39cd483a4199ea945392b8e9616733c3f3e7375395b0
3
  size 14654464
neuronxcc-2.21.33363.0+82129205/MODULE_50933980a498d1588ed1+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_50933980a498d1588ed1+a02c3a36/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_50933980a498d1588ed1+a02c3a36/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cdaf565435be8facd9149bb0db421ca3ab498e5ca8e7ce76c4ac70d52b842ace
3
+ size 654916
neuronxcc-2.21.33363.0+82129205/MODULE_50933980a498d1588ed1+a02c3a36/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:980900bb3a186094e23dfcba022299fb8eceee3b1cf7eaeddb7a3ae04acd96a2
3
+ size 1414144
neuronxcc-2.21.33363.0+82129205/MODULE_50933980a498d1588ed1+a02c3a36/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af0f4b0f2efcf9c3db829e6fd12518adb9bbe53762e248778e139165b7b89f79
3
+ size 1525019
neuronxcc-2.21.33363.0+82129205/MODULE_60f37ef2cb181ebeb434+a02c3a36/model.neff CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1fe4ebd73ffc01c6a9f0e1b592025b6316c1795432ee900302da206df867cc79
3
  size 1772544
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:663e05b29d30cbd492c1f74adc46dca314605eb2658b0f4dcdb5f97f715116f8
3
  size 1772544
neuronxcc-2.21.33363.0+82129205/MODULE_60f37ef2cb181ebeb434+a02c3a36/wrapped_neff.hlo CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9ebfa95057eb5713aa05862ae992aaa77ce4045ea5f806dcc3726efca85a605a
3
  size 1894933
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b1d4da75119b89a223ada9b8876044c4c911d18b2fdbaa6f412a4293bbf5449
3
  size 1894933
neuronxcc-2.21.33363.0+82129205/MODULE_95b9072f246645f24461+a02c3a36/model.hlo_module.pb CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b364b15f88cdba27d3ac4973f7dcfb2cbc49da9abd971bc1f0c52315364d55dc
3
  size 451319
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12bcc80b2ad2cb1e12abbca61cd3f324f6f4f564cf78d121d0f5fcd01305c683
3
  size 451319
neuronxcc-2.21.33363.0+82129205/MODULE_95b9072f246645f24461+a02c3a36/model.neff CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f173e6c1f419b379767e720c82df3bec9c7942c49ca6ce29905fdea0ee240610
3
  size 2509824
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bdf04a70eb8bda2f7f4e49bedb2c346136c7e9b0dec32fa7421ea72474563fbe
3
  size 2509824
neuronxcc-2.21.33363.0+82129205/MODULE_95b9072f246645f24461+a02c3a36/wrapped_neff.hlo CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5cde9a5e1a781062038054f87a06889f4898ef6a5b5a50ae52638bd62e04342a
3
  size 2583911
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c84f3a91584fb71a18f3ba16dd9c1cde9917ff28ab72e432ab03a61ad722f8dc
3
  size 2583911
neuronxcc-2.21.33363.0+82129205/MODULE_a1cd38756c74f5816b8b+24129607/model.hlo_module.pb CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5afb2944ba0711a4441c137bd39995accbbe79dc510f590407ae77b49a33df0d
3
  size 1050679
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db4d29e73bd32d78207490919cc67097223016b8f282251e22eed37a4209717a
3
  size 1050679
neuronxcc-2.21.33363.0+82129205/MODULE_a1cd38756c74f5816b8b+24129607/model.neff CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:652897628aae608aa5ee58fc2ab10b8599eba028f43996e8519fe056c5c402f8
3
  size 5827584
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4807eec7d26a7d456764b6ba7e0004e40c7abd34ae300771c1cfd5cf427aa1d0
3
  size 5827584
neuronxcc-2.21.33363.0+82129205/MODULE_ad16674dd3fc5a51c226+a02c3a36/model.neff CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2f7ccb3c563a4c0fdae234892df68dcd6c60cd41674bf983c7a5693709d2af9c
3
  size 3073024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b8be74d754d3877ff4c9586c81611124a2989d55ff021b4915372155a27fe1a
3
  size 3073024
neuronxcc-2.21.33363.0+82129205/MODULE_ad16674dd3fc5a51c226+a02c3a36/wrapped_neff.hlo CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1a6ac7c0556c0e218f9e95cabc999fa5d73cb79aabcc96ccbc57e3c5325e36e9
3
  size 3147125
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1a2dfe758abc52214d97cbae6fd8bfe738893d16164557e687784ac1c1a80a9
3
  size 3147125
neuronxcc-2.21.33363.0+82129205/MODULE_c69743a3b026cb8c0f9e+a02c3a36/model.hlo_module.pb CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:19abd02589b7e770cc7d22a41832a1d7c1b18e5db1cdbd0f57e95939951b2311
3
  size 588724
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:726697ae6d69106a35fc6caab871c35bccbf129873b887f75986cfbdd687e625
3
  size 588724
neuronxcc-2.21.33363.0+82129205/MODULE_c69743a3b026cb8c0f9e+a02c3a36/model.neff CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9e9a16db24435f3344d68819315ac43e7ec035fdf41c68ba517491e1cc7db394
3
  size 1659904
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a7c258ba5a15e6c435b0719ac65ac7aed62119f5d3d3dc497bc8557452b5afc7
3
  size 1659904
neuronxcc-2.21.33363.0+82129205/MODULE_c69743a3b026cb8c0f9e+a02c3a36/wrapped_neff.hlo CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ee0ea35ca5a7f2dda442d215bb13c572d23bc55b448a956b0f57522e111c3d11
3
  size 1782293
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd30f302fc03f3e2cbf90294931da3ae0f5faa5a5f0cf021340988ef20175088
3
  size 1782293