Hugging Face's logo

aws-neuron
/
optimum-neuron-cache

Model card Files Community
474
dacorvo HF Staff commited on
Commit
7fd08c9
·
verified ·
1 Parent(s): a92f1d6

Synchronizing local compiler cache.

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +13 -0
  2. neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev5/llama/unsloth/Llama-3.2-1B-Instruct/05f1cd1b9a81ce4544f9.json +78 -0
  3. neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev7/llama/llamafactory/tiny-random-Llama-3/007ed0c0cab705897799.json +77 -0
  4. neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev7/llama/llamafactory/tiny-random-Llama-3/892d8ca0f2425da9c03b.json +77 -0
  5. neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev7/llama/llamafactory/tiny-random-Llama-3/a1adaee75c9e8cc04831.json +77 -0
  6. neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev7/llama/llamafactory/tiny-random-Llama-3/a2e466575a68f3e72707.json +77 -0
  7. neuronxcc-2.17.194.0+d312836f/MODULE_06f737474468cb436640+bfe5714b/compile_flags.json +1 -0
  8. neuronxcc-2.17.194.0+d312836f/MODULE_06f737474468cb436640+bfe5714b/model.done +0 -0
  9. neuronxcc-2.17.194.0+d312836f/MODULE_06f737474468cb436640+bfe5714b/model.hlo_module.pb +3 -0
  10. neuronxcc-2.17.194.0+d312836f/MODULE_06f737474468cb436640+bfe5714b/model.neff +3 -0
  11. neuronxcc-2.17.194.0+d312836f/MODULE_228e41176dd02608ed1e+431f5505/compile_flags.json +1 -0
  12. neuronxcc-2.17.194.0+d312836f/MODULE_228e41176dd02608ed1e+431f5505/model.done +0 -0
  13. neuronxcc-2.17.194.0+d312836f/MODULE_228e41176dd02608ed1e+431f5505/model.hlo_module.pb +3 -0
  14. neuronxcc-2.17.194.0+d312836f/MODULE_228e41176dd02608ed1e+431f5505/model.neff +0 -0
  15. neuronxcc-2.17.194.0+d312836f/MODULE_331276a07386ee77d52e+431f5505/compile_flags.json +1 -0
  16. neuronxcc-2.17.194.0+d312836f/MODULE_331276a07386ee77d52e+431f5505/model.done +0 -0
  17. neuronxcc-2.17.194.0+d312836f/MODULE_331276a07386ee77d52e+431f5505/model.hlo_module.pb +3 -0
  18. neuronxcc-2.17.194.0+d312836f/MODULE_331276a07386ee77d52e+431f5505/model.neff +0 -0
  19. neuronxcc-2.17.194.0+d312836f/MODULE_4334b3c1a396e8ac4a6f+165e9558/compile_flags.json +1 -0
  20. neuronxcc-2.17.194.0+d312836f/MODULE_4334b3c1a396e8ac4a6f+165e9558/model.done +0 -0
  21. neuronxcc-2.17.194.0+d312836f/MODULE_4334b3c1a396e8ac4a6f+165e9558/model.hlo_module.pb +3 -0
  22. neuronxcc-2.17.194.0+d312836f/MODULE_4334b3c1a396e8ac4a6f+165e9558/model.neff +3 -0
  23. neuronxcc-2.17.194.0+d312836f/MODULE_4334b3c1a396e8ac4a6f+165e9558/wrapped_neff.hlo +3 -0
  24. neuronxcc-2.17.194.0+d312836f/MODULE_45da94458d038d09c818+613edded/model.neff +1 -1
  25. neuronxcc-2.17.194.0+d312836f/MODULE_55c3972a85a9fd509c79+165e9558/compile_flags.json +1 -0
  26. neuronxcc-2.17.194.0+d312836f/MODULE_55c3972a85a9fd509c79+165e9558/model.done +0 -0
  27. neuronxcc-2.17.194.0+d312836f/MODULE_55c3972a85a9fd509c79+165e9558/model.hlo_module.pb +3 -0
  28. neuronxcc-2.17.194.0+d312836f/MODULE_55c3972a85a9fd509c79+165e9558/model.neff +3 -0
  29. neuronxcc-2.17.194.0+d312836f/MODULE_55c3972a85a9fd509c79+165e9558/wrapped_neff.hlo +3 -0
  30. neuronxcc-2.17.194.0+d312836f/MODULE_6e3263be13428364a15b+bfe5714b/compile_flags.json +1 -0
  31. neuronxcc-2.17.194.0+d312836f/MODULE_6e3263be13428364a15b+bfe5714b/model.done +0 -0
  32. neuronxcc-2.17.194.0+d312836f/MODULE_6e3263be13428364a15b+bfe5714b/model.hlo_module.pb +3 -0
  33. neuronxcc-2.17.194.0+d312836f/MODULE_6e3263be13428364a15b+bfe5714b/model.neff +3 -0
  34. neuronxcc-2.17.194.0+d312836f/MODULE_7d0c7e211138c3505935+165e9558/compile_flags.json +1 -0
  35. neuronxcc-2.17.194.0+d312836f/MODULE_7d0c7e211138c3505935+165e9558/model.done +0 -0
  36. neuronxcc-2.17.194.0+d312836f/MODULE_7d0c7e211138c3505935+165e9558/model.hlo_module.pb +3 -0
  37. neuronxcc-2.17.194.0+d312836f/MODULE_7d0c7e211138c3505935+165e9558/model.neff +3 -0
  38. neuronxcc-2.17.194.0+d312836f/MODULE_7d0c7e211138c3505935+165e9558/wrapped_neff.hlo +3 -0
  39. neuronxcc-2.17.194.0+d312836f/MODULE_81cb45fde7e941e6558d+613edded/model.neff +1 -1
  40. neuronxcc-2.17.194.0+d312836f/MODULE_9cc304c3ad910d31a27e+bfe5714b/compile_flags.json +1 -0
  41. neuronxcc-2.17.194.0+d312836f/MODULE_9cc304c3ad910d31a27e+bfe5714b/model.done +0 -0
  42. neuronxcc-2.17.194.0+d312836f/MODULE_9cc304c3ad910d31a27e+bfe5714b/model.hlo_module.pb +3 -0
  43. neuronxcc-2.17.194.0+d312836f/MODULE_9cc304c3ad910d31a27e+bfe5714b/model.neff +3 -0
  44. neuronxcc-2.17.194.0+d312836f/MODULE_a1d42723f4b151570208+bfe5714b/compile_flags.json +1 -0
  45. neuronxcc-2.17.194.0+d312836f/MODULE_a1d42723f4b151570208+bfe5714b/model.done +0 -0
  46. neuronxcc-2.17.194.0+d312836f/MODULE_a1d42723f4b151570208+bfe5714b/model.hlo_module.pb +3 -0
  47. neuronxcc-2.17.194.0+d312836f/MODULE_a1d42723f4b151570208+bfe5714b/model.neff +3 -0
  48. neuronxcc-2.17.194.0+d312836f/MODULE_a669c4bb07e1a9435b7c+431f5505/compile_flags.json +1 -0
  49. neuronxcc-2.17.194.0+d312836f/MODULE_a669c4bb07e1a9435b7c+431f5505/model.done +0 -0
  50. neuronxcc-2.17.194.0+d312836f/MODULE_a669c4bb07e1a9435b7c+431f5505/model.hlo_module.pb +3 -0
.gitattributes CHANGED
@@ -8240,3 +8240,16 @@ neuronxcc-2.17.194.0+d312836f/MODULE_a30822a1b2d1fbcee30c+bfe5714b/model.neff fi
8240
  neuronxcc-2.17.194.0+d312836f/MODULE_a849442b615562b13ba6+431f5505/model.neff filter=lfs diff=lfs merge=lfs -text
8241
  neuronxcc-2.17.194.0+d312836f/MODULE_ccfcad18cf4ad9b6e4bc+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
8242
  neuronxcc-2.17.194.0+d312836f/MODULE_ccfcad18cf4ad9b6e4bc+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8240
  neuronxcc-2.17.194.0+d312836f/MODULE_a849442b615562b13ba6+431f5505/model.neff filter=lfs diff=lfs merge=lfs -text
8241
  neuronxcc-2.17.194.0+d312836f/MODULE_ccfcad18cf4ad9b6e4bc+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
8242
  neuronxcc-2.17.194.0+d312836f/MODULE_ccfcad18cf4ad9b6e4bc+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
8243
+ neuronxcc-2.17.194.0+d312836f/MODULE_06f737474468cb436640+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
8244
+ neuronxcc-2.17.194.0+d312836f/MODULE_4334b3c1a396e8ac4a6f+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
8245
+ neuronxcc-2.17.194.0+d312836f/MODULE_4334b3c1a396e8ac4a6f+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
8246
+ neuronxcc-2.17.194.0+d312836f/MODULE_55c3972a85a9fd509c79+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
8247
+ neuronxcc-2.17.194.0+d312836f/MODULE_55c3972a85a9fd509c79+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
8248
+ neuronxcc-2.17.194.0+d312836f/MODULE_6e3263be13428364a15b+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
8249
+ neuronxcc-2.17.194.0+d312836f/MODULE_7d0c7e211138c3505935+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
8250
+ neuronxcc-2.17.194.0+d312836f/MODULE_7d0c7e211138c3505935+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
8251
+ neuronxcc-2.17.194.0+d312836f/MODULE_9cc304c3ad910d31a27e+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
8252
+ neuronxcc-2.17.194.0+d312836f/MODULE_a1d42723f4b151570208+bfe5714b/model.neff filter=lfs diff=lfs merge=lfs -text
8253
+ neuronxcc-2.17.194.0+d312836f/MODULE_c2f195a5af02904a4378+431f5505/model.neff filter=lfs diff=lfs merge=lfs -text
8254
+ neuronxcc-2.17.194.0+d312836f/MODULE_fdd1d8e95911bdcc8dec+165e9558/model.neff filter=lfs diff=lfs merge=lfs -text
8255
+ neuronxcc-2.17.194.0+d312836f/MODULE_fdd1d8e95911bdcc8dec+165e9558/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev5/llama/unsloth/Llama-3.2-1B-Instruct/05f1cd1b9a81ce4544f9.json ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "head_dim": 64,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 2048,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 8192,
15
+ "max_position_embeddings": 131072,
16
+ "mlp_bias": false,
17
+ "model_type": "llama",
18
+ "neuron": {
19
+ "_serialized_key": "NxDNeuronConfig",
20
+ "async_mode": false,
21
+ "attn_kernel_enabled": false,
22
+ "batch_size": 4,
23
+ "capacity_factor": null,
24
+ "cc_pipeline_tiling_factor": 2,
25
+ "checkpoint_id": "unsloth/Llama-3.2-1B-Instruct",
26
+ "checkpoint_revision": "5a8abab4a5d6f164389b1079fb721cfab8d7126c",
27
+ "continuous_batching": true,
28
+ "enable_bucketing": false,
29
+ "ep_degree": 1,
30
+ "flash_decoding_enabled": false,
31
+ "fused_qkv": true,
32
+ "glu_mlp": true,
33
+ "is_chunked_prefill": false,
34
+ "local_ranks_size": 24,
35
+ "logical_nc_config": 1,
36
+ "max_batch_size": 4,
37
+ "max_context_length": 4096,
38
+ "max_topk": 256,
39
+ "mlp_kernel_enabled": false,
40
+ "mlp_kernel_fuse_residual_add": false,
41
+ "n_active_tokens": 4096,
42
+ "neuronxcc_version": "2.17.194.0+d312836f",
43
+ "num_cores_per_group": 1,
44
+ "on_device_sampling": true,
45
+ "optimum_neuron_version": "0.2.0.dev5",
46
+ "output_logits": false,
47
+ "padding_side": "right",
48
+ "pp_degree": 1,
49
+ "qk_layernorm": false,
50
+ "qkv_kernel_enabled": false,
51
+ "rpl_reduce_dtype": "bfloat16",
52
+ "sequence_length": 4096,
53
+ "sequence_parallel_enabled": false,
54
+ "speculation_length": 0,
55
+ "start_rank_id": 0,
56
+ "target": null,
57
+ "torch_dtype": "bfloat16",
58
+ "tp_degree": 24,
59
+ "vocab_parallel": false
60
+ },
61
+ "num_attention_heads": 32,
62
+ "num_hidden_layers": 16,
63
+ "num_key_value_heads": 8,
64
+ "pretraining_tp": 1,
65
+ "rms_norm_eps": 1e-05,
66
+ "rope_scaling": {
67
+ "factor": 32.0,
68
+ "high_freq_factor": 4.0,
69
+ "low_freq_factor": 1.0,
70
+ "original_max_position_embeddings": 8192,
71
+ "rope_type": "llama3"
72
+ },
73
+ "rope_theta": 500000.0,
74
+ "tie_word_embeddings": true,
75
+ "unsloth_fixed": true,
76
+ "use_cache": true,
77
+ "vocab_size": 128256
78
+ }
neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev7/llama/llamafactory/tiny-random-Llama-3/007ed0c0cab705897799.json ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "head_dim": 4,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 16,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 64,
15
+ "max_position_embeddings": 131072,
16
+ "mlp_bias": false,
17
+ "model_type": "llama",
18
+ "neuron": {
19
+ "_serialized_key": "NxDNeuronConfig",
20
+ "async_mode": false,
21
+ "attn_kernel_enabled": false,
22
+ "batch_size": 1,
23
+ "capacity_factor": null,
24
+ "cc_pipeline_tiling_factor": 2,
25
+ "checkpoint_id": "llamafactory/tiny-random-Llama-3",
26
+ "checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
27
+ "continuous_batching": false,
28
+ "enable_bucketing": false,
29
+ "ep_degree": 1,
30
+ "flash_decoding_enabled": false,
31
+ "fused_qkv": true,
32
+ "glu_mlp": true,
33
+ "is_chunked_prefill": false,
34
+ "local_ranks_size": 2,
35
+ "logical_nc_config": 1,
36
+ "max_batch_size": 1,
37
+ "max_context_length": 100,
38
+ "max_topk": 256,
39
+ "mlp_kernel_enabled": false,
40
+ "mlp_kernel_fuse_residual_add": false,
41
+ "n_active_tokens": 100,
42
+ "neuronxcc_version": "2.17.194.0+d312836f",
43
+ "num_cores_per_group": 1,
44
+ "on_device_sampling": true,
45
+ "optimum_neuron_version": "0.2.0.dev7",
46
+ "output_logits": false,
47
+ "padding_side": "right",
48
+ "pp_degree": 1,
49
+ "qk_layernorm": false,
50
+ "qkv_kernel_enabled": false,
51
+ "rpl_reduce_dtype": "float16",
52
+ "sequence_length": 100,
53
+ "sequence_parallel_enabled": false,
54
+ "speculation_length": 0,
55
+ "start_rank_id": 0,
56
+ "target": null,
57
+ "torch_dtype": "float16",
58
+ "tp_degree": 2,
59
+ "vocab_parallel": false
60
+ },
61
+ "num_attention_heads": 4,
62
+ "num_hidden_layers": 2,
63
+ "num_key_value_heads": 4,
64
+ "pretraining_tp": 1,
65
+ "rms_norm_eps": 1e-05,
66
+ "rope_scaling": {
67
+ "factor": 8.0,
68
+ "high_freq_factor": 4.0,
69
+ "low_freq_factor": 1.0,
70
+ "original_max_position_embeddings": 8192,
71
+ "rope_type": "llama3"
72
+ },
73
+ "rope_theta": 500000.0,
74
+ "tie_word_embeddings": false,
75
+ "use_cache": true,
76
+ "vocab_size": 128256
77
+ }
neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev7/llama/llamafactory/tiny-random-Llama-3/892d8ca0f2425da9c03b.json ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "head_dim": 4,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 16,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 64,
15
+ "max_position_embeddings": 131072,
16
+ "mlp_bias": false,
17
+ "model_type": "llama",
18
+ "neuron": {
19
+ "_serialized_key": "NxDNeuronConfig",
20
+ "async_mode": false,
21
+ "attn_kernel_enabled": false,
22
+ "batch_size": 1,
23
+ "capacity_factor": null,
24
+ "cc_pipeline_tiling_factor": 2,
25
+ "checkpoint_id": "llamafactory/tiny-random-Llama-3",
26
+ "checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
27
+ "continuous_batching": false,
28
+ "enable_bucketing": false,
29
+ "ep_degree": 1,
30
+ "flash_decoding_enabled": false,
31
+ "fused_qkv": true,
32
+ "glu_mlp": true,
33
+ "is_chunked_prefill": false,
34
+ "local_ranks_size": 2,
35
+ "logical_nc_config": 1,
36
+ "max_batch_size": 1,
37
+ "max_context_length": 100,
38
+ "max_topk": 256,
39
+ "mlp_kernel_enabled": false,
40
+ "mlp_kernel_fuse_residual_add": false,
41
+ "n_active_tokens": 100,
42
+ "neuronxcc_version": "2.17.194.0+d312836f",
43
+ "num_cores_per_group": 1,
44
+ "on_device_sampling": true,
45
+ "optimum_neuron_version": "0.2.0.dev7",
46
+ "output_logits": false,
47
+ "padding_side": "right",
48
+ "pp_degree": 1,
49
+ "qk_layernorm": false,
50
+ "qkv_kernel_enabled": false,
51
+ "rpl_reduce_dtype": "bfloat16",
52
+ "sequence_length": 100,
53
+ "sequence_parallel_enabled": false,
54
+ "speculation_length": 0,
55
+ "start_rank_id": 0,
56
+ "target": null,
57
+ "torch_dtype": "bfloat16",
58
+ "tp_degree": 2,
59
+ "vocab_parallel": false
60
+ },
61
+ "num_attention_heads": 4,
62
+ "num_hidden_layers": 2,
63
+ "num_key_value_heads": 4,
64
+ "pretraining_tp": 1,
65
+ "rms_norm_eps": 1e-05,
66
+ "rope_scaling": {
67
+ "factor": 8.0,
68
+ "high_freq_factor": 4.0,
69
+ "low_freq_factor": 1.0,
70
+ "original_max_position_embeddings": 8192,
71
+ "rope_type": "llama3"
72
+ },
73
+ "rope_theta": 500000.0,
74
+ "tie_word_embeddings": false,
75
+ "use_cache": true,
76
+ "vocab_size": 128256
77
+ }
neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev7/llama/llamafactory/tiny-random-Llama-3/a1adaee75c9e8cc04831.json ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "head_dim": 4,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 16,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 64,
15
+ "max_position_embeddings": 131072,
16
+ "mlp_bias": false,
17
+ "model_type": "llama",
18
+ "neuron": {
19
+ "_serialized_key": "NxDNeuronConfig",
20
+ "async_mode": false,
21
+ "attn_kernel_enabled": false,
22
+ "batch_size": 2,
23
+ "capacity_factor": null,
24
+ "cc_pipeline_tiling_factor": 2,
25
+ "checkpoint_id": "llamafactory/tiny-random-Llama-3",
26
+ "checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
27
+ "continuous_batching": true,
28
+ "enable_bucketing": false,
29
+ "ep_degree": 1,
30
+ "flash_decoding_enabled": false,
31
+ "fused_qkv": true,
32
+ "glu_mlp": true,
33
+ "is_chunked_prefill": false,
34
+ "local_ranks_size": 2,
35
+ "logical_nc_config": 1,
36
+ "max_batch_size": 2,
37
+ "max_context_length": 100,
38
+ "max_topk": 256,
39
+ "mlp_kernel_enabled": false,
40
+ "mlp_kernel_fuse_residual_add": false,
41
+ "n_active_tokens": 100,
42
+ "neuronxcc_version": "2.17.194.0+d312836f",
43
+ "num_cores_per_group": 1,
44
+ "on_device_sampling": false,
45
+ "optimum_neuron_version": "0.2.0.dev7",
46
+ "output_logits": false,
47
+ "padding_side": "right",
48
+ "pp_degree": 1,
49
+ "qk_layernorm": false,
50
+ "qkv_kernel_enabled": false,
51
+ "rpl_reduce_dtype": "float16",
52
+ "sequence_length": 100,
53
+ "sequence_parallel_enabled": false,
54
+ "speculation_length": 0,
55
+ "start_rank_id": 0,
56
+ "target": null,
57
+ "torch_dtype": "float16",
58
+ "tp_degree": 2,
59
+ "vocab_parallel": false
60
+ },
61
+ "num_attention_heads": 4,
62
+ "num_hidden_layers": 2,
63
+ "num_key_value_heads": 4,
64
+ "pretraining_tp": 1,
65
+ "rms_norm_eps": 1e-05,
66
+ "rope_scaling": {
67
+ "factor": 8.0,
68
+ "high_freq_factor": 4.0,
69
+ "low_freq_factor": 1.0,
70
+ "original_max_position_embeddings": 8192,
71
+ "rope_type": "llama3"
72
+ },
73
+ "rope_theta": 500000.0,
74
+ "tie_word_embeddings": false,
75
+ "use_cache": true,
76
+ "vocab_size": 128256
77
+ }
neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev7/llama/llamafactory/tiny-random-Llama-3/a2e466575a68f3e72707.json ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "head_dim": 4,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 16,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 64,
15
+ "max_position_embeddings": 131072,
16
+ "mlp_bias": false,
17
+ "model_type": "llama",
18
+ "neuron": {
19
+ "_serialized_key": "NxDNeuronConfig",
20
+ "async_mode": false,
21
+ "attn_kernel_enabled": false,
22
+ "batch_size": 1,
23
+ "capacity_factor": null,
24
+ "cc_pipeline_tiling_factor": 2,
25
+ "checkpoint_id": "llamafactory/tiny-random-Llama-3",
26
+ "checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
27
+ "continuous_batching": false,
28
+ "enable_bucketing": false,
29
+ "ep_degree": 1,
30
+ "flash_decoding_enabled": false,
31
+ "fused_qkv": true,
32
+ "glu_mlp": true,
33
+ "is_chunked_prefill": false,
34
+ "local_ranks_size": 2,
35
+ "logical_nc_config": 1,
36
+ "max_batch_size": 1,
37
+ "max_context_length": 131072,
38
+ "max_topk": 256,
39
+ "mlp_kernel_enabled": false,
40
+ "mlp_kernel_fuse_residual_add": false,
41
+ "n_active_tokens": 131072,
42
+ "neuronxcc_version": "2.17.194.0+d312836f",
43
+ "num_cores_per_group": 1,
44
+ "on_device_sampling": true,
45
+ "optimum_neuron_version": "0.2.0.dev7",
46
+ "output_logits": false,
47
+ "padding_side": "right",
48
+ "pp_degree": 1,
49
+ "qk_layernorm": false,
50
+ "qkv_kernel_enabled": false,
51
+ "rpl_reduce_dtype": "bfloat16",
52
+ "sequence_length": 131072,
53
+ "sequence_parallel_enabled": false,
54
+ "speculation_length": 0,
55
+ "start_rank_id": 0,
56
+ "target": null,
57
+ "torch_dtype": "bfloat16",
58
+ "tp_degree": 2,
59
+ "vocab_parallel": false
60
+ },
61
+ "num_attention_heads": 4,
62
+ "num_hidden_layers": 2,
63
+ "num_key_value_heads": 4,
64
+ "pretraining_tp": 1,
65
+ "rms_norm_eps": 1e-05,
66
+ "rope_scaling": {
67
+ "factor": 8.0,
68
+ "high_freq_factor": 4.0,
69
+ "low_freq_factor": 1.0,
70
+ "original_max_position_embeddings": 8192,
71
+ "rope_type": "llama3"
72
+ },
73
+ "rope_theta": 500000.0,
74
+ "tie_word_embeddings": false,
75
+ "use_cache": true,
76
+ "vocab_size": 128256
77
+ }
neuronxcc-2.17.194.0+d312836f/MODULE_06f737474468cb436640+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_06f737474468cb436640+bfe5714b/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_06f737474468cb436640+bfe5714b/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a902c97b9db908e03187b21a6c01701c50bfe0df9f6b71b809da7f4f16af05c9
3
+ size 80831
neuronxcc-2.17.194.0+d312836f/MODULE_06f737474468cb436640+bfe5714b/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af0f5381043c9ef05027aa98ac3c13eb4414875c5b4fcdec7b813f04c3df912b
3
+ size 226304
neuronxcc-2.17.194.0+d312836f/MODULE_228e41176dd02608ed1e+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_228e41176dd02608ed1e+431f5505/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_228e41176dd02608ed1e+431f5505/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e23ddd2c1ceb0ca9b13f12cfe085496dda4a9903f08aa51dd4aaa763af0ab3e
3
+ size 7106
neuronxcc-2.17.194.0+d312836f/MODULE_228e41176dd02608ed1e+431f5505/model.neff ADDED
Binary file (42 kB). View file
 
neuronxcc-2.17.194.0+d312836f/MODULE_331276a07386ee77d52e+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_331276a07386ee77d52e+431f5505/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_331276a07386ee77d52e+431f5505/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f32d5db6cdf19c397941133ad17626fbb60335e60d3115b19046ba4d38e111f0
3
+ size 7106
neuronxcc-2.17.194.0+d312836f/MODULE_331276a07386ee77d52e+431f5505/model.neff ADDED
Binary file (42 kB). View file
 
neuronxcc-2.17.194.0+d312836f/MODULE_4334b3c1a396e8ac4a6f+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_4334b3c1a396e8ac4a6f+165e9558/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_4334b3c1a396e8ac4a6f+165e9558/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d938d24775001bb3a9c7fe290d4f0c9459eee3f32d275af31c99748fd81e81f
3
+ size 423227
neuronxcc-2.17.194.0+d312836f/MODULE_4334b3c1a396e8ac4a6f+165e9558/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62f8c764353e2db9bb07fb9dc00b13fedf5cf29339faa18bf684bbbf5b19c272
3
+ size 2806784
neuronxcc-2.17.194.0+d312836f/MODULE_4334b3c1a396e8ac4a6f+165e9558/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:382ceeb24e11646e57fb436c0d64fbb0ce59caa919bd5f1a92d3d35308ed4a00
3
+ size 2875313
neuronxcc-2.17.194.0+d312836f/MODULE_45da94458d038d09c818+613edded/model.neff CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:abfc595163761b975d91957ba7e7e0a12dda49a12bd80321ca9643fbaa9caadc
3
  size 18371584
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:05390870b066c993234ecd93a81cf6032e9775f84fefd55396ea230a4f453a0e
3
  size 18371584
neuronxcc-2.17.194.0+d312836f/MODULE_55c3972a85a9fd509c79+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_55c3972a85a9fd509c79+165e9558/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_55c3972a85a9fd509c79+165e9558/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4b5d0271ec9be39982fe3aeb4554340968a29392c77f6a65b122a47790efeb9
3
+ size 80798
neuronxcc-2.17.194.0+d312836f/MODULE_55c3972a85a9fd509c79+165e9558/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e5bf722c4149f80ae71df74bae7079026871b56527cedd7a875d34831ea33292
3
+ size 236544
neuronxcc-2.17.194.0+d312836f/MODULE_55c3972a85a9fd509c79+165e9558/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a377b14bff0e1489f2e43ba5d0d3123c57d434accfc9481612f9935106a758a8
3
+ size 244319
neuronxcc-2.17.194.0+d312836f/MODULE_6e3263be13428364a15b+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_6e3263be13428364a15b+bfe5714b/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_6e3263be13428364a15b+bfe5714b/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b4d3726f2d8c619c2a146cfb9d0b4b7589f1a0de7bd8a2438446f534cd3a0f18
3
+ size 91018
neuronxcc-2.17.194.0+d312836f/MODULE_6e3263be13428364a15b+bfe5714b/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a5e24e8b14ca4aabf4f19a5c7e081e69613c617752ef631fbde8cba746b5e08
3
+ size 213853184
neuronxcc-2.17.194.0+d312836f/MODULE_7d0c7e211138c3505935+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_7d0c7e211138c3505935+165e9558/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_7d0c7e211138c3505935+165e9558/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:170784eb623597dcaedd6e649d1cf8a28a6ed509fa2a9789be34c9c4fa4e5828
3
+ size 80988
neuronxcc-2.17.194.0+d312836f/MODULE_7d0c7e211138c3505935+165e9558/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2730e31a5dc351c9439641fe4ea14db3e69d1113ca90cd381dc33e6891d67813
3
+ size 461824
neuronxcc-2.17.194.0+d312836f/MODULE_7d0c7e211138c3505935+165e9558/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:36ac7c6a3343116727f7006d8c68267b9644dc27ead59c6c24b3de5c6ab1a703
3
+ size 469663
neuronxcc-2.17.194.0+d312836f/MODULE_81cb45fde7e941e6558d+613edded/model.neff CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1f76308b46268c61640b1d0f61fcfaaa5f1f30dcc7a4cf125c3cc70f1a94e4ac
3
  size 1414144
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:75ba441f792a712d68d09aebc1d574b9a301bb19ef05fccac5daf7202cb44137
3
  size 1414144
neuronxcc-2.17.194.0+d312836f/MODULE_9cc304c3ad910d31a27e+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_9cc304c3ad910d31a27e+bfe5714b/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_9cc304c3ad910d31a27e+bfe5714b/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5abb3fdf792bb77cd45770a5d89fd5b633ebd23393f46ea7f65742c18e235d12
3
+ size 493167
neuronxcc-2.17.194.0+d312836f/MODULE_9cc304c3ad910d31a27e+bfe5714b/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:50111b7b7febc7d24fc401a6e23e2596e480fa09e8393964bf01c6ac067afdc7
3
+ size 5407744
neuronxcc-2.17.194.0+d312836f/MODULE_a1d42723f4b151570208+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_a1d42723f4b151570208+bfe5714b/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_a1d42723f4b151570208+bfe5714b/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89e0e274b15f298d10619f21c5cba58d538fc37e1547bfd401d8c6064c09b709
3
+ size 80831
neuronxcc-2.17.194.0+d312836f/MODULE_a1d42723f4b151570208+bfe5714b/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b11be33c81d54d95b9ff736e03c0987c8436a9b5f06872d2fe4ea6d0ef9b612f
3
+ size 226304
neuronxcc-2.17.194.0+d312836f/MODULE_a669c4bb07e1a9435b7c+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_a669c4bb07e1a9435b7c+431f5505/model.done ADDED
File without changes
neuronxcc-2.17.194.0+d312836f/MODULE_a669c4bb07e1a9435b7c+431f5505/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:01d4291445f0fad2b3267c9fc11713f66b969c4b7def78292de1e0ef96a0c206
3
+ size 7106