Hugging Face's logo

aws-neuron
/
optimum-neuron-cache

Model card Files
xet
Community
630
dacorvo HF Staff commited on
Commit
16e28ed
·
verified ·
1 Parent(s): 02dd753

Synchronizing local compiler cache.

Browse files
Files changed (17) hide show
  1. .gitattributes +4 -0
  2. neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.3.0.dev1/qwen2/Qwen/Qwen2.5-72B-Instruct/b2a8c0e4fbc830460cd8.json +70 -0
  3. neuronxcc-2.17.194.0+d312836f/MODULE_76278e8d8ac0ba75d5b8+165e9558/compile_flags.json +1 -0
  4. neuronxcc-2.17.194.0+d312836f/MODULE_76278e8d8ac0ba75d5b8+165e9558/model.done +0 -0
  5. neuronxcc-2.17.194.0+d312836f/MODULE_76278e8d8ac0ba75d5b8+165e9558/model.hlo_module.pb +3 -0
  6. neuronxcc-2.17.194.0+d312836f/MODULE_76278e8d8ac0ba75d5b8+165e9558/model.neff +3 -0
  7. neuronxcc-2.17.194.0+d312836f/MODULE_76278e8d8ac0ba75d5b8+165e9558/wrapped_neff.hlo +3 -0
  8. neuronxcc-2.17.194.0+d312836f/MODULE_7d6793c439ef9de8c1e3+431f5505/compile_flags.json +1 -0
  9. neuronxcc-2.17.194.0+d312836f/MODULE_7d6793c439ef9de8c1e3+431f5505/model.done +0 -0
  10. neuronxcc-2.17.194.0+d312836f/MODULE_7d6793c439ef9de8c1e3+431f5505/model.hlo_module.pb +3 -0
  11. neuronxcc-2.17.194.0+d312836f/MODULE_7d6793c439ef9de8c1e3+431f5505/model.neff +3 -0
  12. neuronxcc-2.17.194.0+d312836f/MODULE_89b3018566aae05ef5ba+bfe5714b/compile_flags.json +1 -0
  13. neuronxcc-2.17.194.0+d312836f/MODULE_89b3018566aae05ef5ba+bfe5714b/model.done +0 -0
  14. neuronxcc-2.17.194.0+d312836f/MODULE_89b3018566aae05ef5ba+bfe5714b/model.hlo_module.pb +3 -0
  15. neuronxcc-2.17.194.0+d312836f/MODULE_89b3018566aae05ef5ba+bfe5714b/model.neff +3 -0
  16. neuronxcc-2.17.194.0+d312836f/MODULE_8ad4d0dd4ea9d9c2284a+165e9558/compile_flags.json +1 -0
  17. neuronxcc-2.17.194.0+d312836f/MODULE_8ad4d0dd4ea9d9c2284a+165e9558/model.hlo_module.pb +3 -0
.gitattributes CHANGED
@@ -8555,3 +8555,7 @@ neuronxcc-2.17.194.0+d312836f/MODULE_d632e0868986304fe3d6+165e9558/wrapped_neff.
8555
  neuronxcc-2.17.194.0+d312836f/MODULE_03f21958f3e3c04efe30+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
8556
  neuronxcc-2.17.194.0+d312836f/MODULE_03f21958f3e3c04efe30+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
8557
  neuronxcc-2.17.194.0+d312836f/MODULE_a4678584537b90a372ac+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
8555
  neuronxcc-2.17.194.0+d312836f/MODULE_03f21958f3e3c04efe30+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
8556
  neuronxcc-2.17.194.0+d312836f/MODULE_03f21958f3e3c04efe30+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
8557
  neuronxcc-2.17.194.0+d312836f/MODULE_a4678584537b90a372ac+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
8558
+ neuronxcc-2.17.194.0+d312836f/MODULE_76278e8d8ac0ba75d5b8+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
8559
+ neuronxcc-2.17.194.0+d312836f/MODULE_76278e8d8ac0ba75d5b8+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
8560
+ neuronxcc-2.17.194.0+d312836f/MODULE_7d6793c439ef9de8c1e3+431f5505/model.neff filter=lfs diff=lfs merge=lfs -text
8561
+ neuronxcc-2.17.194.0+d312836f/MODULE_89b3018566aae05ef5ba+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.3.0.dev1/qwen2/Qwen/Qwen2.5-72B-Instruct/b2a8c0e4fbc830460cd8.json ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen2.5-72B-Instruct",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen2ForCausalLM"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 8192,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 29568,
13
+ "max_position_embeddings": 32768,
14
+ "max_window_layers": 70,
15
+ "model_type": "qwen2",
16
+ "neuron": {
17
+ "_serialized_key": "NxDNeuronConfig",
18
+ "async_mode": false,
19
+ "attn_kernel_enabled": false,
20
+ "batch_size": 4,
21
+ "capacity_factor": null,
22
+ "cc_pipeline_tiling_factor": 2,
23
+ "checkpoint_id": "Qwen/Qwen2.5-72B-Instruct",
24
+ "checkpoint_revision": "495f39366efef23836d0cfae4fbe635880d2be31",
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": 24,
33
+ "logical_nc_config": 1,
34
+ "max_batch_size": 4,
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": 24,
57
+ "vocab_parallel": false
58
+ },
59
+ "num_attention_heads": 64,
60
+ "num_hidden_layers": 80,
61
+ "num_key_value_heads": 8,
62
+ "rms_norm_eps": 1e-06,
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_76278e8d8ac0ba75d5b8+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_76278e8d8ac0ba75d5b8+165e9558/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_76278e8d8ac0ba75d5b8+165e9558/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a87386a00f65cab9ad146a114f0d19eaeb76951cb6bb7f960e57b50d07296a75
3
+ size 2025265
neuronxcc-2.17.194.0+d312836f/MODULE_76278e8d8ac0ba75d5b8+165e9558/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:adb43c6152134d2cdcbca0d98e69c0a090ed134f586fc8ee541238b57b092e45
3
+ size 10363904
neuronxcc-2.17.194.0+d312836f/MODULE_76278e8d8ac0ba75d5b8+165e9558/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a00d926a237813c39aa00c9e46fbd753bbfdc97a8ff9fe8b33a0fa2c4e7bef3e
3
+ size 10856578
neuronxcc-2.17.194.0+d312836f/MODULE_7d6793c439ef9de8c1e3+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_7d6793c439ef9de8c1e3+431f5505/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_7d6793c439ef9de8c1e3+431f5505/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:982f176ce2424d1e9c79f44f49543c308fd91f14cc11fc4b8fae598d554bf067
3
+ size 511803
neuronxcc-2.17.194.0+d312836f/MODULE_7d6793c439ef9de8c1e3+431f5505/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ba152daeb549fb27400ca57062cf08ccc0393fe25cf145c22b2a1cbaf0ec7740
3
+ size 7855104
neuronxcc-2.17.194.0+d312836f/MODULE_89b3018566aae05ef5ba+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_89b3018566aae05ef5ba+bfe5714b/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_89b3018566aae05ef5ba+bfe5714b/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bff634d0a755283e97a782d1b86afbd2a8c32b588c800463bd112c2af3839ccb
3
+ size 2421840
neuronxcc-2.17.194.0+d312836f/MODULE_89b3018566aae05ef5ba+bfe5714b/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4fa79d2d9d9d2fd2b8f09e53cb6b93a2cf197dd266393db1e968a84585799126
3
+ size 3544064
neuronxcc-2.17.194.0+d312836f/MODULE_8ad4d0dd4ea9d9c2284a+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_8ad4d0dd4ea9d9c2284a+165e9558/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7066ffbe82f115bbe274eab1fb37356e731a69b70a68c3dbbe1af4af0240e422
3
+ size 2054666