Synchronizing local compiler cache.
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- .gitattributes +15 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/943a2aeabe8c8143be7f.json +58 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama/llamafactory/tiny-random-Llama-3/1f52dd56ba4b6c4c1f74.json +62 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama/llamafactory/tiny-random-Llama-3/e453ca73065785a3cca2.json +62 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama/llamafactory/tiny-random-Llama-3/eb5e22c6ea67b0c94017.json +62 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama/unsloth/Llama-3.2-1B-Instruct/8b999d87128760e77f02.json +63 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama/unsloth/Llama-3.2-1B-Instruct/c4f37eab4a64f353cc38.json +63 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama4_text/tiny-random/llama-4/4a8f7bc879bccb57966e.json +81 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/mixtral/dacorvo/Mixtral-tiny/48f4a9ff0be5e8420a1e.json +58 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/phi3/yujiepan/phi-4-tiny-random/b12b7ef8a6c0d7ba0ec0.json +59 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen2/Qwen/Qwen2.5-0.5B/9e4a267147d9ad7df5ef.json +82 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen2/Qwen/Qwen2.5-0.5B/a00506c78d9c3ead4d9d.json +82 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen2/Qwen/Qwen2.5-0.5B/b52a2f6aa7b5f8e5bb5a.json +82 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen2/yujiepan/qwen2.5-128k-tiny-random/3bb0bffc287e06003671.json +64 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/bc617080acd95b57b216.json +87 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/bd1f95813da05a700637.json +87 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/385c39cc1e6ad9337849.json +65 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/smollm3/HuggingFaceTB/SmolLM3-3B/1dd41ec7acb6511daeb0.json +134 -0
- neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/smollm3/HuggingFaceTB/SmolLM3-3B/e0ff29206a1e88178465.json +134 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_039fa65bf5a4ef68f834+08b9b048/compile_flags.json +1 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_039fa65bf5a4ef68f834+08b9b048/model.done +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_039fa65bf5a4ef68f834+08b9b048/model.hlo_module.pb +3 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_039fa65bf5a4ef68f834+08b9b048/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_10217061096959125489+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_10244305442015770634+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_10606948783918825529+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_10645643398657092095+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_10661660426924300837+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_10746122569655005679+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_11086318750207148626+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_11909310600244571805+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_11933818254123612383+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_1210392327607194823+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_12487216553200321032+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_13085549342645515693+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_13102010590082783346+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_13475273426270778455+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_13728813963059599796+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_13783872956654054643+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_14097776130612710282+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_14155105943057125178+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_14173682194645188821+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_14274833313744358348+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_14469716438804935215+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_14739672283771660808+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_15442663025941492357+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_15557971880365771457+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_15564425602929126510+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_15589838460896944293+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.18209.0+043b1bf7/MODULE_15654572849171857535+e30acd3a/model.neff +0 -0
.gitattributes
CHANGED
|
@@ -13017,3 +13017,18 @@ neuronxcc-2.21.18209.0+043b1bf7/MODULE_5c83b0bb7257da8a674c+9cb2fe70/model.neff
|
|
| 13017 |
neuronxcc-2.21.18209.0+043b1bf7/MODULE_5c83b0bb7257da8a674c+9cb2fe70/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13018 |
neuronxcc-2.21.18209.0+043b1bf7/MODULE_b10fffb847021ecf80eb+562c86ea/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13019 |
neuronxcc-2.21.18209.0+043b1bf7/MODULE_b10fffb847021ecf80eb+562c86ea/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13017 |
neuronxcc-2.21.18209.0+043b1bf7/MODULE_5c83b0bb7257da8a674c+9cb2fe70/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13018 |
neuronxcc-2.21.18209.0+043b1bf7/MODULE_b10fffb847021ecf80eb+562c86ea/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13019 |
neuronxcc-2.21.18209.0+043b1bf7/MODULE_b10fffb847021ecf80eb+562c86ea/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13020 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_285442f0c8e665e1c464+fdadbfb3/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13021 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_285442f0c8e665e1c464+fdadbfb3/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13022 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_2a2a38ff534d69dee9d8+1a5ed1fb/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13023 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_2a2a38ff534d69dee9d8+1a5ed1fb/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13024 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_2d53e7d37242178f5553+384bca3f/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13025 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_2d53e7d37242178f5553+384bca3f/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13026 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_7e6ec8464ad7c3c6284f+b91a68d9/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13027 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_7e6ec8464ad7c3c6284f+b91a68d9/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13028 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_aaa9f7d6fac20c922790+e4ec13ff/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13029 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_aaa9f7d6fac20c922790+e4ec13ff/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13030 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_b3cfe28275e1a02a015c+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13031 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_d4df6648fd68fe444e68+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13032 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_d4df6648fd68fe444e68+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13033 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_d5115466754a2222278e+246d8cec/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13034 |
+
neuronxcc-2.21.18209.0+043b1bf7/MODULE_d5115466754a2222278e+246d8cec/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/943a2aeabe8c8143be7f.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"GraniteForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"attention_multiplier": 1.0,
|
| 11 |
+
"dtype": "float32",
|
| 12 |
+
"embedding_multiplier": 1.0,
|
| 13 |
+
"hidden_act": "silu",
|
| 14 |
+
"hidden_size": 32,
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 64,
|
| 17 |
+
"logits_scaling": 1.0,
|
| 18 |
+
"max_position_embeddings": 2048,
|
| 19 |
+
"mlp_bias": false,
|
| 20 |
+
"model_type": "granite",
|
| 21 |
+
"neuron": {
|
| 22 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 23 |
+
"batch_size": 1,
|
| 24 |
+
"capacity_factor": null,
|
| 25 |
+
"checkpoint_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
|
| 26 |
+
"checkpoint_revision": "c3074ebc0ac2fe545305f5e5f6cce2cc9b2aa0c5",
|
| 27 |
+
"continuous_batching": false,
|
| 28 |
+
"ep_degree": 1,
|
| 29 |
+
"fused_qkv": true,
|
| 30 |
+
"glu_mlp": true,
|
| 31 |
+
"local_ranks_size": 2,
|
| 32 |
+
"max_batch_size": 1,
|
| 33 |
+
"max_context_length": 1024,
|
| 34 |
+
"max_topk": 256,
|
| 35 |
+
"n_active_tokens": 1024,
|
| 36 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 37 |
+
"on_device_sampling": true,
|
| 38 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 39 |
+
"output_logits": false,
|
| 40 |
+
"pp_degree": 1,
|
| 41 |
+
"sequence_length": 1024,
|
| 42 |
+
"speculation_length": 0,
|
| 43 |
+
"start_rank_id": 0,
|
| 44 |
+
"target": "trn1",
|
| 45 |
+
"torch_dtype": "float32",
|
| 46 |
+
"tp_degree": 2
|
| 47 |
+
},
|
| 48 |
+
"num_attention_heads": 4,
|
| 49 |
+
"num_hidden_layers": 2,
|
| 50 |
+
"num_key_value_heads": 4,
|
| 51 |
+
"residual_multiplier": 1.0,
|
| 52 |
+
"rms_norm_eps": 1e-06,
|
| 53 |
+
"rope_scaling": null,
|
| 54 |
+
"rope_theta": 10000.0,
|
| 55 |
+
"tie_word_embeddings": false,
|
| 56 |
+
"use_cache": true,
|
| 57 |
+
"vocab_size": 49152
|
| 58 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama/llamafactory/tiny-random-Llama-3/1f52dd56ba4b6c4c1f74.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "llamafactory/tiny-random-Llama-3",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"LlamaForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"dtype": "float16",
|
| 11 |
+
"head_dim": 4,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 16,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 64,
|
| 16 |
+
"max_position_embeddings": 131072,
|
| 17 |
+
"mlp_bias": false,
|
| 18 |
+
"model_type": "llama",
|
| 19 |
+
"neuron": {
|
| 20 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 21 |
+
"batch_size": 1,
|
| 22 |
+
"capacity_factor": null,
|
| 23 |
+
"checkpoint_id": "llamafactory/tiny-random-Llama-3",
|
| 24 |
+
"checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
|
| 25 |
+
"continuous_batching": false,
|
| 26 |
+
"ep_degree": 1,
|
| 27 |
+
"fused_qkv": true,
|
| 28 |
+
"glu_mlp": true,
|
| 29 |
+
"local_ranks_size": 2,
|
| 30 |
+
"max_batch_size": 1,
|
| 31 |
+
"max_context_length": 1024,
|
| 32 |
+
"max_topk": 256,
|
| 33 |
+
"n_active_tokens": 1024,
|
| 34 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 35 |
+
"on_device_sampling": true,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 1024,
|
| 40 |
+
"speculation_length": 0,
|
| 41 |
+
"start_rank_id": 0,
|
| 42 |
+
"target": "trn1",
|
| 43 |
+
"torch_dtype": "float16",
|
| 44 |
+
"tp_degree": 2
|
| 45 |
+
},
|
| 46 |
+
"num_attention_heads": 4,
|
| 47 |
+
"num_hidden_layers": 2,
|
| 48 |
+
"num_key_value_heads": 4,
|
| 49 |
+
"pretraining_tp": 1,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_scaling": {
|
| 52 |
+
"factor": 8.0,
|
| 53 |
+
"high_freq_factor": 4.0,
|
| 54 |
+
"low_freq_factor": 1.0,
|
| 55 |
+
"original_max_position_embeddings": 8192,
|
| 56 |
+
"rope_type": "llama3"
|
| 57 |
+
},
|
| 58 |
+
"rope_theta": 500000.0,
|
| 59 |
+
"tie_word_embeddings": false,
|
| 60 |
+
"use_cache": true,
|
| 61 |
+
"vocab_size": 128256
|
| 62 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama/llamafactory/tiny-random-Llama-3/e453ca73065785a3cca2.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": 512,
|
| 32 |
+
"max_topk": 256,
|
| 33 |
+
"n_active_tokens": 512,
|
| 34 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 35 |
+
"on_device_sampling": true,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 512,
|
| 40 |
+
"speculation_length": 0,
|
| 41 |
+
"start_rank_id": 0,
|
| 42 |
+
"target": "trn1",
|
| 43 |
+
"torch_dtype": "float16",
|
| 44 |
+
"tp_degree": 2
|
| 45 |
+
},
|
| 46 |
+
"num_attention_heads": 4,
|
| 47 |
+
"num_hidden_layers": 2,
|
| 48 |
+
"num_key_value_heads": 4,
|
| 49 |
+
"pretraining_tp": 1,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_scaling": {
|
| 52 |
+
"factor": 8.0,
|
| 53 |
+
"high_freq_factor": 4.0,
|
| 54 |
+
"low_freq_factor": 1.0,
|
| 55 |
+
"original_max_position_embeddings": 8192,
|
| 56 |
+
"rope_type": "llama3"
|
| 57 |
+
},
|
| 58 |
+
"rope_theta": 500000.0,
|
| 59 |
+
"tie_word_embeddings": false,
|
| 60 |
+
"use_cache": true,
|
| 61 |
+
"vocab_size": 128256
|
| 62 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama/llamafactory/tiny-random-Llama-3/eb5e22c6ea67b0c94017.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": 2,
|
| 22 |
+
"capacity_factor": null,
|
| 23 |
+
"checkpoint_id": "llamafactory/tiny-random-Llama-3",
|
| 24 |
+
"checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
|
| 25 |
+
"continuous_batching": true,
|
| 26 |
+
"ep_degree": 1,
|
| 27 |
+
"fused_qkv": true,
|
| 28 |
+
"glu_mlp": true,
|
| 29 |
+
"local_ranks_size": 2,
|
| 30 |
+
"max_batch_size": 2,
|
| 31 |
+
"max_context_length": 128,
|
| 32 |
+
"max_topk": 256,
|
| 33 |
+
"n_active_tokens": 128,
|
| 34 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 35 |
+
"on_device_sampling": true,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 128,
|
| 40 |
+
"speculation_length": 0,
|
| 41 |
+
"start_rank_id": 0,
|
| 42 |
+
"target": "trn2",
|
| 43 |
+
"torch_dtype": "float16",
|
| 44 |
+
"tp_degree": 2
|
| 45 |
+
},
|
| 46 |
+
"num_attention_heads": 4,
|
| 47 |
+
"num_hidden_layers": 2,
|
| 48 |
+
"num_key_value_heads": 4,
|
| 49 |
+
"pretraining_tp": 1,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_scaling": {
|
| 52 |
+
"factor": 8.0,
|
| 53 |
+
"high_freq_factor": 4.0,
|
| 54 |
+
"low_freq_factor": 1.0,
|
| 55 |
+
"original_max_position_embeddings": 8192,
|
| 56 |
+
"rope_type": "llama3"
|
| 57 |
+
},
|
| 58 |
+
"rope_theta": 500000.0,
|
| 59 |
+
"tie_word_embeddings": false,
|
| 60 |
+
"use_cache": true,
|
| 61 |
+
"vocab_size": 128256
|
| 62 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama/unsloth/Llama-3.2-1B-Instruct/8b999d87128760e77f02.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": "2.21.18209.0+043b1bf7",
|
| 35 |
+
"on_device_sampling": false,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 4096,
|
| 40 |
+
"speculation_length": 5,
|
| 41 |
+
"start_rank_id": 0,
|
| 42 |
+
"target": "trn1",
|
| 43 |
+
"torch_dtype": "bfloat16",
|
| 44 |
+
"tp_degree": 2
|
| 45 |
+
},
|
| 46 |
+
"num_attention_heads": 32,
|
| 47 |
+
"num_hidden_layers": 16,
|
| 48 |
+
"num_key_value_heads": 8,
|
| 49 |
+
"pretraining_tp": 1,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_scaling": {
|
| 52 |
+
"factor": 32.0,
|
| 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": true,
|
| 60 |
+
"unsloth_fixed": true,
|
| 61 |
+
"use_cache": true,
|
| 62 |
+
"vocab_size": 128256
|
| 63 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama/unsloth/Llama-3.2-1B-Instruct/c4f37eab4a64f353cc38.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": "2.21.18209.0+043b1bf7",
|
| 35 |
+
"on_device_sampling": false,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 4096,
|
| 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": 32,
|
| 47 |
+
"num_hidden_layers": 16,
|
| 48 |
+
"num_key_value_heads": 8,
|
| 49 |
+
"pretraining_tp": 1,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_scaling": {
|
| 52 |
+
"factor": 32.0,
|
| 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": true,
|
| 60 |
+
"unsloth_fixed": true,
|
| 61 |
+
"use_cache": true,
|
| 62 |
+
"vocab_size": 128256
|
| 63 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama4_text/tiny-random/llama-4/4a8f7bc879bccb57966e.json
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "tiny-random/llama-4",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_chunk_size": 128,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attn_scale": 0.1,
|
| 9 |
+
"attn_temperature_tuning": 4,
|
| 10 |
+
"cache_implementation": "hybrid",
|
| 11 |
+
"dtype": "bfloat16",
|
| 12 |
+
"floor_scale": 8192,
|
| 13 |
+
"for_llm_compressor": false,
|
| 14 |
+
"head_dim": 32,
|
| 15 |
+
"hidden_act": "silu",
|
| 16 |
+
"hidden_size": 32,
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"interleave_moe_layer_step": 2,
|
| 19 |
+
"intermediate_size": 64,
|
| 20 |
+
"intermediate_size_mlp": 128,
|
| 21 |
+
"layer_types": [
|
| 22 |
+
"chunked_attention",
|
| 23 |
+
"chunked_attention",
|
| 24 |
+
"chunked_attention",
|
| 25 |
+
"full_attention"
|
| 26 |
+
],
|
| 27 |
+
"max_position_embeddings": 1048576,
|
| 28 |
+
"model_type": "llama4_text",
|
| 29 |
+
"moe_layers": [
|
| 30 |
+
1,
|
| 31 |
+
3
|
| 32 |
+
],
|
| 33 |
+
"neuron": {
|
| 34 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 35 |
+
"batch_size": 1,
|
| 36 |
+
"capacity_factor": null,
|
| 37 |
+
"checkpoint_id": "tiny-random/llama-4",
|
| 38 |
+
"checkpoint_revision": "9e716f5d4d1ffe0a44a15f46f4a12b840439aba4",
|
| 39 |
+
"continuous_batching": false,
|
| 40 |
+
"ep_degree": 1,
|
| 41 |
+
"fused_qkv": false,
|
| 42 |
+
"glu_mlp": true,
|
| 43 |
+
"local_ranks_size": 2,
|
| 44 |
+
"max_batch_size": 1,
|
| 45 |
+
"max_context_length": 1024,
|
| 46 |
+
"max_topk": 256,
|
| 47 |
+
"n_active_tokens": 1024,
|
| 48 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 49 |
+
"on_device_sampling": true,
|
| 50 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 51 |
+
"output_logits": false,
|
| 52 |
+
"pp_degree": 1,
|
| 53 |
+
"sequence_length": 1024,
|
| 54 |
+
"speculation_length": 0,
|
| 55 |
+
"start_rank_id": 0,
|
| 56 |
+
"target": "trn1",
|
| 57 |
+
"torch_dtype": "bfloat16",
|
| 58 |
+
"tp_degree": 2
|
| 59 |
+
},
|
| 60 |
+
"no_rope_layers": [
|
| 61 |
+
1,
|
| 62 |
+
1,
|
| 63 |
+
1,
|
| 64 |
+
0
|
| 65 |
+
],
|
| 66 |
+
"num_attention_heads": 1,
|
| 67 |
+
"num_experts_per_tok": 1,
|
| 68 |
+
"num_hidden_layers": 4,
|
| 69 |
+
"num_key_value_heads": 1,
|
| 70 |
+
"num_local_experts": 8,
|
| 71 |
+
"output_router_logits": false,
|
| 72 |
+
"rms_norm_eps": 1e-05,
|
| 73 |
+
"rope_scaling": null,
|
| 74 |
+
"rope_theta": 500000.0,
|
| 75 |
+
"router_aux_loss_coef": 0.001,
|
| 76 |
+
"router_jitter_noise": 0.0,
|
| 77 |
+
"tie_word_embeddings": true,
|
| 78 |
+
"use_cache": true,
|
| 79 |
+
"use_qk_norm": true,
|
| 80 |
+
"vocab_size": 202048
|
| 81 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/mixtral/dacorvo/Mixtral-tiny/48f4a9ff0be5e8420a1e.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "dacorvo/Mixtral-tiny",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"MixtralForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"dtype": "float16",
|
| 10 |
+
"head_dim": 32,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 1024,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 3584,
|
| 15 |
+
"max_position_embeddings": 1024,
|
| 16 |
+
"model_type": "mixtral",
|
| 17 |
+
"neuron": {
|
| 18 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 19 |
+
"batch_size": 1,
|
| 20 |
+
"capacity_factor": null,
|
| 21 |
+
"checkpoint_id": "dacorvo/Mixtral-tiny",
|
| 22 |
+
"checkpoint_revision": "c557ba205ddff6ea911f4719e0d543d6c08356b6",
|
| 23 |
+
"continuous_batching": false,
|
| 24 |
+
"ep_degree": 1,
|
| 25 |
+
"fused_qkv": false,
|
| 26 |
+
"glu_mlp": true,
|
| 27 |
+
"local_ranks_size": 2,
|
| 28 |
+
"max_batch_size": 1,
|
| 29 |
+
"max_context_length": 1024,
|
| 30 |
+
"max_topk": 256,
|
| 31 |
+
"n_active_tokens": 1024,
|
| 32 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 33 |
+
"on_device_sampling": false,
|
| 34 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 35 |
+
"output_logits": false,
|
| 36 |
+
"pp_degree": 1,
|
| 37 |
+
"sequence_length": 1024,
|
| 38 |
+
"speculation_length": 0,
|
| 39 |
+
"start_rank_id": 0,
|
| 40 |
+
"target": "trn1",
|
| 41 |
+
"torch_dtype": "float16",
|
| 42 |
+
"tp_degree": 2
|
| 43 |
+
},
|
| 44 |
+
"num_attention_heads": 32,
|
| 45 |
+
"num_experts_per_tok": 2,
|
| 46 |
+
"num_hidden_layers": 2,
|
| 47 |
+
"num_key_value_heads": 8,
|
| 48 |
+
"num_local_experts": 8,
|
| 49 |
+
"output_router_logits": false,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_theta": 10000.0,
|
| 52 |
+
"router_aux_loss_coef": 0.001,
|
| 53 |
+
"router_jitter_noise": 0.0,
|
| 54 |
+
"sliding_window": 4096,
|
| 55 |
+
"tie_word_embeddings": false,
|
| 56 |
+
"use_cache": true,
|
| 57 |
+
"vocab_size": 32000
|
| 58 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/phi3/yujiepan/phi-4-tiny-random/b12b7ef8a6c0d7ba0ec0.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "yujiepan/phi-4-tiny-random",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Phi3ForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"auto_map": {},
|
| 11 |
+
"dtype": "bfloat16",
|
| 12 |
+
"embd_pdrop": 0.0,
|
| 13 |
+
"hidden_act": "silu",
|
| 14 |
+
"hidden_size": 16,
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 32,
|
| 17 |
+
"max_position_embeddings": 16384,
|
| 18 |
+
"model_type": "phi3",
|
| 19 |
+
"neuron": {
|
| 20 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 21 |
+
"batch_size": 1,
|
| 22 |
+
"capacity_factor": null,
|
| 23 |
+
"checkpoint_id": "yujiepan/phi-4-tiny-random",
|
| 24 |
+
"checkpoint_revision": "18a9a1168dc97ac6d128f811925670c275610f5a",
|
| 25 |
+
"continuous_batching": false,
|
| 26 |
+
"ep_degree": 1,
|
| 27 |
+
"fused_qkv": true,
|
| 28 |
+
"glu_mlp": true,
|
| 29 |
+
"local_ranks_size": 2,
|
| 30 |
+
"max_batch_size": 1,
|
| 31 |
+
"max_context_length": 1024,
|
| 32 |
+
"max_topk": 256,
|
| 33 |
+
"n_active_tokens": 1024,
|
| 34 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 35 |
+
"on_device_sampling": true,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 1024,
|
| 40 |
+
"speculation_length": 0,
|
| 41 |
+
"start_rank_id": 0,
|
| 42 |
+
"target": "trn1",
|
| 43 |
+
"torch_dtype": "bfloat16",
|
| 44 |
+
"tp_degree": 2
|
| 45 |
+
},
|
| 46 |
+
"num_attention_heads": 2,
|
| 47 |
+
"num_hidden_layers": 2,
|
| 48 |
+
"num_key_value_heads": 1,
|
| 49 |
+
"original_max_position_embeddings": 16384,
|
| 50 |
+
"partial_rotary_factor": 1.0,
|
| 51 |
+
"resid_pdrop": 0.0,
|
| 52 |
+
"rms_norm_eps": 1e-05,
|
| 53 |
+
"rope_scaling": null,
|
| 54 |
+
"rope_theta": 250000,
|
| 55 |
+
"sliding_window": null,
|
| 56 |
+
"tie_word_embeddings": false,
|
| 57 |
+
"use_cache": true,
|
| 58 |
+
"vocab_size": 100352
|
| 59 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen2/Qwen/Qwen2.5-0.5B/9e4a267147d9ad7df5ef.json
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": 1,
|
| 46 |
+
"capacity_factor": null,
|
| 47 |
+
"checkpoint_id": "Qwen/Qwen2.5-0.5B",
|
| 48 |
+
"checkpoint_revision": "060db6499f32faf8b98477b0a26969ef7d8b9987",
|
| 49 |
+
"continuous_batching": false,
|
| 50 |
+
"ep_degree": 1,
|
| 51 |
+
"fused_qkv": false,
|
| 52 |
+
"glu_mlp": true,
|
| 53 |
+
"local_ranks_size": 2,
|
| 54 |
+
"max_batch_size": 1,
|
| 55 |
+
"max_context_length": 4096,
|
| 56 |
+
"max_topk": 256,
|
| 57 |
+
"n_active_tokens": 4096,
|
| 58 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 59 |
+
"on_device_sampling": true,
|
| 60 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 61 |
+
"output_logits": false,
|
| 62 |
+
"pp_degree": 1,
|
| 63 |
+
"sequence_length": 4096,
|
| 64 |
+
"speculation_length": 0,
|
| 65 |
+
"start_rank_id": 0,
|
| 66 |
+
"target": "trn1",
|
| 67 |
+
"torch_dtype": "bfloat16",
|
| 68 |
+
"tp_degree": 2
|
| 69 |
+
},
|
| 70 |
+
"num_attention_heads": 14,
|
| 71 |
+
"num_hidden_layers": 24,
|
| 72 |
+
"num_key_value_heads": 2,
|
| 73 |
+
"rms_norm_eps": 1e-06,
|
| 74 |
+
"rope_scaling": null,
|
| 75 |
+
"rope_theta": 1000000.0,
|
| 76 |
+
"sliding_window": null,
|
| 77 |
+
"tie_word_embeddings": true,
|
| 78 |
+
"use_cache": true,
|
| 79 |
+
"use_mrope": false,
|
| 80 |
+
"use_sliding_window": false,
|
| 81 |
+
"vocab_size": 151936
|
| 82 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen2/Qwen/Qwen2.5-0.5B/a00506c78d9c3ead4d9d.json
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": 4,
|
| 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": 4,
|
| 55 |
+
"max_context_length": 4096,
|
| 56 |
+
"max_topk": 256,
|
| 57 |
+
"n_active_tokens": 4096,
|
| 58 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 59 |
+
"on_device_sampling": false,
|
| 60 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 61 |
+
"output_logits": false,
|
| 62 |
+
"pp_degree": 1,
|
| 63 |
+
"sequence_length": 4096,
|
| 64 |
+
"speculation_length": 0,
|
| 65 |
+
"start_rank_id": 0,
|
| 66 |
+
"target": "trn1",
|
| 67 |
+
"torch_dtype": "bfloat16",
|
| 68 |
+
"tp_degree": 2
|
| 69 |
+
},
|
| 70 |
+
"num_attention_heads": 14,
|
| 71 |
+
"num_hidden_layers": 24,
|
| 72 |
+
"num_key_value_heads": 2,
|
| 73 |
+
"rms_norm_eps": 1e-06,
|
| 74 |
+
"rope_scaling": null,
|
| 75 |
+
"rope_theta": 1000000.0,
|
| 76 |
+
"sliding_window": null,
|
| 77 |
+
"tie_word_embeddings": true,
|
| 78 |
+
"use_cache": true,
|
| 79 |
+
"use_mrope": false,
|
| 80 |
+
"use_sliding_window": false,
|
| 81 |
+
"vocab_size": 151936
|
| 82 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen2/Qwen/Qwen2.5-0.5B/b52a2f6aa7b5f8e5bb5a.json
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": "2.21.18209.0+043b1bf7",
|
| 59 |
+
"on_device_sampling": false,
|
| 60 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 61 |
+
"output_logits": false,
|
| 62 |
+
"pp_degree": 1,
|
| 63 |
+
"sequence_length": 1024,
|
| 64 |
+
"speculation_length": 0,
|
| 65 |
+
"start_rank_id": 0,
|
| 66 |
+
"target": "trn1",
|
| 67 |
+
"torch_dtype": "bfloat16",
|
| 68 |
+
"tp_degree": 2
|
| 69 |
+
},
|
| 70 |
+
"num_attention_heads": 14,
|
| 71 |
+
"num_hidden_layers": 24,
|
| 72 |
+
"num_key_value_heads": 2,
|
| 73 |
+
"rms_norm_eps": 1e-06,
|
| 74 |
+
"rope_scaling": null,
|
| 75 |
+
"rope_theta": 1000000.0,
|
| 76 |
+
"sliding_window": null,
|
| 77 |
+
"tie_word_embeddings": true,
|
| 78 |
+
"use_cache": true,
|
| 79 |
+
"use_mrope": false,
|
| 80 |
+
"use_sliding_window": false,
|
| 81 |
+
"vocab_size": 151936
|
| 82 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen2/yujiepan/qwen2.5-128k-tiny-random/3bb0bffc287e06003671.json
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "yujiepan/qwen2.5-128k-tiny-random",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Qwen2ForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"hidden_act": "silu",
|
| 11 |
+
"hidden_size": 8,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 16,
|
| 14 |
+
"layer_types": [
|
| 15 |
+
"full_attention",
|
| 16 |
+
"full_attention"
|
| 17 |
+
],
|
| 18 |
+
"max_position_embeddings": 32768,
|
| 19 |
+
"max_window_layers": 1,
|
| 20 |
+
"model_type": "qwen2",
|
| 21 |
+
"neuron": {
|
| 22 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 23 |
+
"batch_size": 1,
|
| 24 |
+
"capacity_factor": null,
|
| 25 |
+
"checkpoint_id": "yujiepan/qwen2.5-128k-tiny-random",
|
| 26 |
+
"checkpoint_revision": "c8296d4ca3f87782876d2382fbb6481d1beb8ef0",
|
| 27 |
+
"continuous_batching": false,
|
| 28 |
+
"ep_degree": 1,
|
| 29 |
+
"fused_qkv": false,
|
| 30 |
+
"glu_mlp": true,
|
| 31 |
+
"local_ranks_size": 2,
|
| 32 |
+
"max_batch_size": 1,
|
| 33 |
+
"max_context_length": 1024,
|
| 34 |
+
"max_topk": 256,
|
| 35 |
+
"n_active_tokens": 1024,
|
| 36 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 37 |
+
"on_device_sampling": true,
|
| 38 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 39 |
+
"output_logits": false,
|
| 40 |
+
"pp_degree": 1,
|
| 41 |
+
"sequence_length": 1024,
|
| 42 |
+
"speculation_length": 0,
|
| 43 |
+
"start_rank_id": 0,
|
| 44 |
+
"target": "trn1",
|
| 45 |
+
"torch_dtype": "bfloat16",
|
| 46 |
+
"tp_degree": 2
|
| 47 |
+
},
|
| 48 |
+
"num_attention_heads": 4,
|
| 49 |
+
"num_hidden_layers": 2,
|
| 50 |
+
"num_key_value_heads": 2,
|
| 51 |
+
"rms_norm_eps": 1e-06,
|
| 52 |
+
"rope_scaling": {
|
| 53 |
+
"factor": 4.0,
|
| 54 |
+
"original_max_position_embeddings": 32768,
|
| 55 |
+
"rope_type": "yarn",
|
| 56 |
+
"type": "yarn"
|
| 57 |
+
},
|
| 58 |
+
"rope_theta": 1000000.0,
|
| 59 |
+
"sliding_window": null,
|
| 60 |
+
"tie_word_embeddings": false,
|
| 61 |
+
"use_cache": true,
|
| 62 |
+
"use_sliding_window": false,
|
| 63 |
+
"vocab_size": 152064
|
| 64 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/bc617080acd95b57b216.json
ADDED
|
@@ -0,0 +1,87 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "Qwen/Qwen3-Embedding-0.6B",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Qwen3ForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"dtype": "bfloat16",
|
| 11 |
+
"head_dim": 128,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 1024,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 3072,
|
| 16 |
+
"layer_types": [
|
| 17 |
+
"full_attention",
|
| 18 |
+
"full_attention",
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention",
|
| 21 |
+
"full_attention",
|
| 22 |
+
"full_attention",
|
| 23 |
+
"full_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"full_attention",
|
| 26 |
+
"full_attention",
|
| 27 |
+
"full_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"full_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention",
|
| 35 |
+
"full_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"full_attention",
|
| 38 |
+
"full_attention",
|
| 39 |
+
"full_attention",
|
| 40 |
+
"full_attention",
|
| 41 |
+
"full_attention",
|
| 42 |
+
"full_attention",
|
| 43 |
+
"full_attention",
|
| 44 |
+
"full_attention"
|
| 45 |
+
],
|
| 46 |
+
"max_position_embeddings": 32768,
|
| 47 |
+
"max_window_layers": 28,
|
| 48 |
+
"model_type": "qwen3",
|
| 49 |
+
"neuron": {
|
| 50 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 51 |
+
"batch_size": 4,
|
| 52 |
+
"capacity_factor": null,
|
| 53 |
+
"checkpoint_id": "Qwen/Qwen3-Embedding-0.6B",
|
| 54 |
+
"checkpoint_revision": "c54f2e6e80b2d7b7de06f51cec4959f6b3e03418",
|
| 55 |
+
"continuous_batching": false,
|
| 56 |
+
"ep_degree": 1,
|
| 57 |
+
"fused_qkv": true,
|
| 58 |
+
"glu_mlp": true,
|
| 59 |
+
"local_ranks_size": 2,
|
| 60 |
+
"max_batch_size": 4,
|
| 61 |
+
"max_context_length": 1024,
|
| 62 |
+
"max_topk": 256,
|
| 63 |
+
"n_active_tokens": 1024,
|
| 64 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 65 |
+
"on_device_sampling": false,
|
| 66 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 67 |
+
"output_logits": false,
|
| 68 |
+
"pp_degree": 1,
|
| 69 |
+
"sequence_length": 1024,
|
| 70 |
+
"speculation_length": 0,
|
| 71 |
+
"start_rank_id": 0,
|
| 72 |
+
"target": "trn1",
|
| 73 |
+
"torch_dtype": "bfloat16",
|
| 74 |
+
"tp_degree": 2
|
| 75 |
+
},
|
| 76 |
+
"num_attention_heads": 16,
|
| 77 |
+
"num_hidden_layers": 28,
|
| 78 |
+
"num_key_value_heads": 8,
|
| 79 |
+
"rms_norm_eps": 1e-06,
|
| 80 |
+
"rope_scaling": null,
|
| 81 |
+
"rope_theta": 1000000,
|
| 82 |
+
"sliding_window": null,
|
| 83 |
+
"tie_word_embeddings": true,
|
| 84 |
+
"use_cache": true,
|
| 85 |
+
"use_sliding_window": false,
|
| 86 |
+
"vocab_size": 151669
|
| 87 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen3/Qwen/Qwen3-Embedding-0.6B/bd1f95813da05a700637.json
ADDED
|
@@ -0,0 +1,87 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "Qwen/Qwen3-Embedding-0.6B",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Qwen3ForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"dtype": "bfloat16",
|
| 11 |
+
"head_dim": 128,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 1024,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 3072,
|
| 16 |
+
"layer_types": [
|
| 17 |
+
"full_attention",
|
| 18 |
+
"full_attention",
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention",
|
| 21 |
+
"full_attention",
|
| 22 |
+
"full_attention",
|
| 23 |
+
"full_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"full_attention",
|
| 26 |
+
"full_attention",
|
| 27 |
+
"full_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"full_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention",
|
| 35 |
+
"full_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"full_attention",
|
| 38 |
+
"full_attention",
|
| 39 |
+
"full_attention",
|
| 40 |
+
"full_attention",
|
| 41 |
+
"full_attention",
|
| 42 |
+
"full_attention",
|
| 43 |
+
"full_attention",
|
| 44 |
+
"full_attention"
|
| 45 |
+
],
|
| 46 |
+
"max_position_embeddings": 32768,
|
| 47 |
+
"max_window_layers": 28,
|
| 48 |
+
"model_type": "qwen3",
|
| 49 |
+
"neuron": {
|
| 50 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 51 |
+
"batch_size": 6,
|
| 52 |
+
"capacity_factor": null,
|
| 53 |
+
"checkpoint_id": "Qwen/Qwen3-Embedding-0.6B",
|
| 54 |
+
"checkpoint_revision": "c54f2e6e80b2d7b7de06f51cec4959f6b3e03418",
|
| 55 |
+
"continuous_batching": false,
|
| 56 |
+
"ep_degree": 1,
|
| 57 |
+
"fused_qkv": true,
|
| 58 |
+
"glu_mlp": true,
|
| 59 |
+
"local_ranks_size": 2,
|
| 60 |
+
"max_batch_size": 6,
|
| 61 |
+
"max_context_length": 1024,
|
| 62 |
+
"max_topk": 256,
|
| 63 |
+
"n_active_tokens": 1024,
|
| 64 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 65 |
+
"on_device_sampling": false,
|
| 66 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 67 |
+
"output_logits": false,
|
| 68 |
+
"pp_degree": 1,
|
| 69 |
+
"sequence_length": 1024,
|
| 70 |
+
"speculation_length": 0,
|
| 71 |
+
"start_rank_id": 0,
|
| 72 |
+
"target": "trn1",
|
| 73 |
+
"torch_dtype": "bfloat16",
|
| 74 |
+
"tp_degree": 2
|
| 75 |
+
},
|
| 76 |
+
"num_attention_heads": 16,
|
| 77 |
+
"num_hidden_layers": 28,
|
| 78 |
+
"num_key_value_heads": 8,
|
| 79 |
+
"rms_norm_eps": 1e-06,
|
| 80 |
+
"rope_scaling": null,
|
| 81 |
+
"rope_theta": 1000000,
|
| 82 |
+
"sliding_window": null,
|
| 83 |
+
"tie_word_embeddings": true,
|
| 84 |
+
"use_cache": true,
|
| 85 |
+
"use_sliding_window": false,
|
| 86 |
+
"vocab_size": 151669
|
| 87 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/385c39cc1e6ad9337849.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "optimum-internal-testing/tiny-random-qwen3_moe",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Qwen3MoeForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"decoder_sparse_step": 2,
|
| 11 |
+
"dtype": "float32",
|
| 12 |
+
"head_dim": 32,
|
| 13 |
+
"hidden_act": "silu",
|
| 14 |
+
"hidden_size": 64,
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 128,
|
| 17 |
+
"max_position_embeddings": 40960,
|
| 18 |
+
"max_window_layers": 1,
|
| 19 |
+
"mlp_only_layers": [],
|
| 20 |
+
"model_type": "qwen3_moe",
|
| 21 |
+
"moe_intermediate_size": 128,
|
| 22 |
+
"neuron": {
|
| 23 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 24 |
+
"batch_size": 1,
|
| 25 |
+
"capacity_factor": null,
|
| 26 |
+
"checkpoint_id": "optimum-internal-testing/tiny-random-qwen3_moe",
|
| 27 |
+
"checkpoint_revision": "e0230be2839556b44b7400a233c73c74b4abb7af",
|
| 28 |
+
"continuous_batching": false,
|
| 29 |
+
"ep_degree": 1,
|
| 30 |
+
"fused_qkv": false,
|
| 31 |
+
"glu_mlp": true,
|
| 32 |
+
"local_ranks_size": 2,
|
| 33 |
+
"max_batch_size": 1,
|
| 34 |
+
"max_context_length": 1024,
|
| 35 |
+
"max_topk": 256,
|
| 36 |
+
"n_active_tokens": 1024,
|
| 37 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 38 |
+
"on_device_sampling": true,
|
| 39 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 40 |
+
"output_logits": false,
|
| 41 |
+
"pp_degree": 1,
|
| 42 |
+
"sequence_length": 1024,
|
| 43 |
+
"speculation_length": 0,
|
| 44 |
+
"start_rank_id": 0,
|
| 45 |
+
"target": "trn1",
|
| 46 |
+
"torch_dtype": "float32",
|
| 47 |
+
"tp_degree": 2
|
| 48 |
+
},
|
| 49 |
+
"norm_topk_prob": true,
|
| 50 |
+
"num_attention_heads": 2,
|
| 51 |
+
"num_experts": 8,
|
| 52 |
+
"num_experts_per_tok": 2,
|
| 53 |
+
"num_hidden_layers": 2,
|
| 54 |
+
"num_key_value_heads": 1,
|
| 55 |
+
"output_router_logits": false,
|
| 56 |
+
"rms_norm_eps": 1e-06,
|
| 57 |
+
"rope_scaling": null,
|
| 58 |
+
"rope_theta": 1000000.0,
|
| 59 |
+
"router_aux_loss_coef": 0.001,
|
| 60 |
+
"sliding_window": null,
|
| 61 |
+
"tie_word_embeddings": true,
|
| 62 |
+
"use_cache": true,
|
| 63 |
+
"use_sliding_window": false,
|
| 64 |
+
"vocab_size": 151936
|
| 65 |
+
}
|
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/smollm3/HuggingFaceTB/SmolLM3-3B/1dd41ec7acb6511daeb0.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": "2.21.18209.0+043b1bf7",
|
| 73 |
+
"on_device_sampling": true,
|
| 74 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 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-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/smollm3/HuggingFaceTB/SmolLM3-3B/e0ff29206a1e88178465.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": 4,
|
| 60 |
+
"capacity_factor": null,
|
| 61 |
+
"checkpoint_id": "HuggingFaceTB/SmolLM3-3B",
|
| 62 |
+
"checkpoint_revision": "a07cc9a04f16550a088caea529712d1d335b0ac1",
|
| 63 |
+
"continuous_batching": true,
|
| 64 |
+
"ep_degree": 1,
|
| 65 |
+
"fused_qkv": true,
|
| 66 |
+
"glu_mlp": true,
|
| 67 |
+
"local_ranks_size": 2,
|
| 68 |
+
"max_batch_size": 4,
|
| 69 |
+
"max_context_length": 4096,
|
| 70 |
+
"max_topk": 256,
|
| 71 |
+
"n_active_tokens": 4096,
|
| 72 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 73 |
+
"on_device_sampling": true,
|
| 74 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 75 |
+
"output_logits": false,
|
| 76 |
+
"pp_degree": 1,
|
| 77 |
+
"sequence_length": 4096,
|
| 78 |
+
"speculation_length": 0,
|
| 79 |
+
"start_rank_id": 0,
|
| 80 |
+
"target": "trn1",
|
| 81 |
+
"torch_dtype": "bfloat16",
|
| 82 |
+
"tp_degree": 2
|
| 83 |
+
},
|
| 84 |
+
"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-2.21.18209.0+043b1bf7/MODULE_039fa65bf5a4ef68f834+08b9b048/compile_flags.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
["--target=trn1", "--enable-saturate-infinity", "--auto-cast=none", "--model-type=transformer", "-O1", "--logfile=/tmp/nxdi_test_2b919496-3f5f-4d6a-a1d6-8227f9a1ee4f/compiler_workdir/Llama3RotaryEmbedding/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_039fa65bf5a4ef68f834+08b9b048/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_039fa65bf5a4ef68f834+08b9b048/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f4a29d4cf27aec9af90e53324eff9ee69e033679482f4016a9dd3e5416c7c68a
|
| 3 |
+
size 1584
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_039fa65bf5a4ef68f834+08b9b048/model.neff
ADDED
|
Binary file (42 kB). View file
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_10217061096959125489+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10217061096959125489+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10217061096959125489+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_10244305442015770634+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10244305442015770634+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10244305442015770634+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_10606948783918825529+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10606948783918825529+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10606948783918825529+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_10645643398657092095+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10645643398657092095+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10645643398657092095+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_10661660426924300837+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10661660426924300837+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10661660426924300837+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_10746122569655005679+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10746122569655005679+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_10746122569655005679+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_11086318750207148626+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_11086318750207148626+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_11086318750207148626+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_11909310600244571805+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_11909310600244571805+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_11909310600244571805+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_11933818254123612383+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_11933818254123612383+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_11933818254123612383+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_1210392327607194823+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_1210392327607194823+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_1210392327607194823+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_12487216553200321032+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_12487216553200321032+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_12487216553200321032+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_13085549342645515693+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_13085549342645515693+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_13085549342645515693+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_13102010590082783346+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_13102010590082783346+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_13102010590082783346+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_13475273426270778455+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_13475273426270778455+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_13475273426270778455+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_13728813963059599796+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_13728813963059599796+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_13728813963059599796+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_13783872956654054643+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_13783872956654054643+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_13783872956654054643+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_14097776130612710282+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14097776130612710282+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14097776130612710282+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_14155105943057125178+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14155105943057125178+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14155105943057125178+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_14173682194645188821+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14173682194645188821+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14173682194645188821+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_14274833313744358348+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14274833313744358348+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14274833313744358348+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_14469716438804935215+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14469716438804935215+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14469716438804935215+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_14739672283771660808+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14739672283771660808+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_14739672283771660808+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_15442663025941492357+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_15442663025941492357+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_15442663025941492357+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_15557971880365771457+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_15557971880365771457+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_15557971880365771457+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_15564425602929126510+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_15564425602929126510+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_15564425602929126510+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_15589838460896944293+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_15589838460896944293+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_15589838460896944293+e30acd3a/model.neff differ
|
|
|
neuronxcc-2.21.18209.0+043b1bf7/MODULE_15654572849171857535+e30acd3a/model.neff
CHANGED
|
Binary files a/neuronxcc-2.21.18209.0+043b1bf7/MODULE_15654572849171857535+e30acd3a/model.neff and b/neuronxcc-2.21.18209.0+043b1bf7/MODULE_15654572849171857535+e30acd3a/model.neff differ
|
|
|