Synchronizing local compiler cache.
Browse files- .gitattributes +4 -0
- neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.3.0.dev1/qwen2/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B/343e6fdcc1a57c199073.json +70 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/compile_flags.json +1 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/model.done +0 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/model.hlo_module.pb +3 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/model.neff +3 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/wrapped_neff.hlo +3 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_51de7bf727a3a094effa+431f5505/compile_flags.json +1 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_51de7bf727a3a094effa+431f5505/model.done +0 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_51de7bf727a3a094effa+431f5505/model.hlo_module.pb +3 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_51de7bf727a3a094effa+431f5505/model.neff +3 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_c808adc36dc289df1d8d+bfe5714b/compile_flags.json +1 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_c808adc36dc289df1d8d+bfe5714b/model.done +0 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_c808adc36dc289df1d8d+bfe5714b/model.hlo_module.pb +3 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_c808adc36dc289df1d8d+bfe5714b/model.neff +3 -0
.gitattributes
CHANGED
@@ -8516,3 +8516,7 @@ neuronxcc-2.17.194.0+d312836f/MODULE_29a9ba12f14b2ee1f5ee+165e9558/model.neff fi
|
|
8516 |
neuronxcc-2.17.194.0+d312836f/MODULE_29a9ba12f14b2ee1f5ee+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
8517 |
neuronxcc-2.17.194.0+d312836f/MODULE_bedc896892ef410d98d5+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
|
8518 |
neuronxcc-2.17.194.0+d312836f/MODULE_e14513207c04240f3c36+431f5505/model.neff filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
8516 |
neuronxcc-2.17.194.0+d312836f/MODULE_29a9ba12f14b2ee1f5ee+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
8517 |
neuronxcc-2.17.194.0+d312836f/MODULE_bedc896892ef410d98d5+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
|
8518 |
neuronxcc-2.17.194.0+d312836f/MODULE_e14513207c04240f3c36+431f5505/model.neff filter=lfs diff=lfs merge=lfs -text
|
8519 |
+
neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
|
8520 |
+
neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
8521 |
+
neuronxcc-2.17.194.0+d312836f/MODULE_51de7bf727a3a094effa+431f5505/model.neff filter=lfs diff=lfs merge=lfs -text
|
8522 |
+
neuronxcc-2.17.194.0+d312836f/MODULE_c808adc36dc289df1d8d+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
|
neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.3.0.dev1/qwen2/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B/343e6fdcc1a57c199073.json
ADDED
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"Qwen2ForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"hidden_act": "silu",
|
10 |
+
"hidden_size": 5120,
|
11 |
+
"initializer_range": 0.02,
|
12 |
+
"intermediate_size": 27648,
|
13 |
+
"max_position_embeddings": 131072,
|
14 |
+
"max_window_layers": 64,
|
15 |
+
"model_type": "qwen2",
|
16 |
+
"neuron": {
|
17 |
+
"_serialized_key": "NxDNeuronConfig",
|
18 |
+
"async_mode": false,
|
19 |
+
"attn_kernel_enabled": false,
|
20 |
+
"batch_size": 8,
|
21 |
+
"capacity_factor": null,
|
22 |
+
"cc_pipeline_tiling_factor": 2,
|
23 |
+
"checkpoint_id": "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B",
|
24 |
+
"checkpoint_revision": "711ad2ea6aa40cfca18895e8aca02ab92df1a746",
|
25 |
+
"continuous_batching": true,
|
26 |
+
"enable_bucketing": false,
|
27 |
+
"ep_degree": 1,
|
28 |
+
"flash_decoding_enabled": false,
|
29 |
+
"fused_qkv": false,
|
30 |
+
"glu_mlp": true,
|
31 |
+
"is_chunked_prefill": false,
|
32 |
+
"local_ranks_size": 8,
|
33 |
+
"logical_nc_config": 1,
|
34 |
+
"max_batch_size": 8,
|
35 |
+
"max_context_length": 4096,
|
36 |
+
"max_topk": 256,
|
37 |
+
"mlp_kernel_enabled": false,
|
38 |
+
"mlp_kernel_fuse_residual_add": false,
|
39 |
+
"n_active_tokens": 4096,
|
40 |
+
"neuronxcc_version": "2.17.194.0+d312836f",
|
41 |
+
"num_cores_per_group": 1,
|
42 |
+
"on_device_sampling": true,
|
43 |
+
"optimum_neuron_version": "0.3.0.dev1",
|
44 |
+
"output_logits": false,
|
45 |
+
"padding_side": "right",
|
46 |
+
"pp_degree": 1,
|
47 |
+
"qk_layernorm": false,
|
48 |
+
"qkv_kernel_enabled": false,
|
49 |
+
"rpl_reduce_dtype": "bfloat16",
|
50 |
+
"sequence_length": 4096,
|
51 |
+
"sequence_parallel_enabled": false,
|
52 |
+
"speculation_length": 0,
|
53 |
+
"start_rank_id": 0,
|
54 |
+
"target": null,
|
55 |
+
"torch_dtype": "bfloat16",
|
56 |
+
"tp_degree": 8,
|
57 |
+
"vocab_parallel": false
|
58 |
+
},
|
59 |
+
"num_attention_heads": 40,
|
60 |
+
"num_hidden_layers": 64,
|
61 |
+
"num_key_value_heads": 8,
|
62 |
+
"rms_norm_eps": 1e-05,
|
63 |
+
"rope_scaling": null,
|
64 |
+
"rope_theta": 1000000.0,
|
65 |
+
"sliding_window": 131072,
|
66 |
+
"tie_word_embeddings": false,
|
67 |
+
"use_cache": true,
|
68 |
+
"use_sliding_window": false,
|
69 |
+
"vocab_size": 152064
|
70 |
+
}
|
neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/compile_flags.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
"--auto-cast=none --model-type=transformer --tensorizer-options='--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-dge-dma --vectorize-strided-dma ' -O2 --internal-num-neuroncores-per-sengine=1 --logfile=/tmp/nxd_model/token_generation_model/_tp0_bk0/log-neuron-cc.txt --enable-internal-neff-wrapper"
|
neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/model.done
ADDED
File without changes
|
neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/model.hlo_module.pb
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0ccc1b873f861174834383c6b11201ee56ce4aa06526dd682a0218485c8a7ad4
|
3 |
+
size 1610010
|
neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/model.neff
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:af7436057e43d6f615eb4b4b5963409876f4409e5acd9e921377f179dc76adf0
|
3 |
+
size 11346944
|
neuronxcc-2.17.194.0+d312836f/MODULE_01abc9376daa952e0718+165e9558/wrapped_neff.hlo
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bc5859ebd491d70702ef04969039c88188fd3ef2571aefb7b35a47e6316b351a
|
3 |
+
size 11742289
|
neuronxcc-2.17.194.0+d312836f/MODULE_51de7bf727a3a094effa+431f5505/compile_flags.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
"--model-type=transformer -O1 --lnc=1 --internal-hlo2tensorizer-options=--experimental-unsafe-fp8e4m3fn-as-fp8e4m3 --logfile=/tmp/nxd_model/layout_opt/log-neuron-cc.txt"
|
neuronxcc-2.17.194.0+d312836f/MODULE_51de7bf727a3a094effa+431f5505/model.done
ADDED
File without changes
|
neuronxcc-2.17.194.0+d312836f/MODULE_51de7bf727a3a094effa+431f5505/model.hlo_module.pb
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ec8fc87fcc95cc82130d02d97d6ff37f75b28795a7c87fdc049022544c7811da
|
3 |
+
size 411300
|
neuronxcc-2.17.194.0+d312836f/MODULE_51de7bf727a3a094effa+431f5505/model.neff
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f02f7d7ce8d684b801799d39e0590fe1eeabf551d594d863b30f53aa7f5c1011
|
3 |
+
size 7261184
|
neuronxcc-2.17.194.0+d312836f/MODULE_c808adc36dc289df1d8d+bfe5714b/compile_flags.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
"--auto-cast=none --model-type=transformer --tensorizer-options='--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-dge-dma --vectorize-strided-dma ' -O2 --internal-num-neuroncores-per-sengine=1 --logfile=/tmp/nxd_model/context_encoding_model/_tp0_bk0/log-neuron-cc.txt"
|
neuronxcc-2.17.194.0+d312836f/MODULE_c808adc36dc289df1d8d+bfe5714b/model.done
ADDED
File without changes
|
neuronxcc-2.17.194.0+d312836f/MODULE_c808adc36dc289df1d8d+bfe5714b/model.hlo_module.pb
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:dfb28582046a4f5d322bf6c69ecde577f961ea505d2ca0860c305d255369fee1
|
3 |
+
size 1928010
|
neuronxcc-2.17.194.0+d312836f/MODULE_c808adc36dc289df1d8d+bfe5714b/model.neff
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:76d63c486e5428cf130623784b922ec1216906a4f354e3bdfe22a61d6227209a
|
3 |
+
size 4660224
|