dacorvo HF Staff commited on
Commit
9693514
·
verified ·
1 Parent(s): b39757c

Synchronizing local compiler cache.

Browse files
.gitattributes CHANGED
@@ -5883,3 +5883,6 @@ neuronxcc-2.21.33363.0+82129205/MODULE_baeb27d758dca75a3243+a02c3a36/wrapped_nef
5883
  neuronxcc-2.21.33363.0+82129205/MODULE_d1e075f89e8272c07272+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
5884
  neuronxcc-2.21.33363.0+82129205/MODULE_d9f17410cfb469b24055+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
5885
  neuronxcc-2.21.33363.0+82129205/MODULE_d9f17410cfb469b24055+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
 
 
 
 
5883
  neuronxcc-2.21.33363.0+82129205/MODULE_d1e075f89e8272c07272+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
5884
  neuronxcc-2.21.33363.0+82129205/MODULE_d9f17410cfb469b24055+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
5885
  neuronxcc-2.21.33363.0+82129205/MODULE_d9f17410cfb469b24055+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
5886
+ neuronxcc-2.21.33363.0+82129205/MODULE_4e5d28a9a05b8ba71697+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
5887
+ neuronxcc-2.21.33363.0+82129205/MODULE_4e5d28a9a05b8ba71697+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
5888
+ neuronxcc-2.21.33363.0+82129205/MODULE_8e336cee3fb428dcae05+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.4.dev1/granite/ibm-granite/granite-3.1-2b-instruct/c7cc8db735ece7aeba17.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "ibm-granite/granite-3.1-2b-instruct",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "GraniteForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.1,
10
+ "attention_multiplier": 0.015625,
11
+ "dtype": "bfloat16",
12
+ "embedding_multiplier": 12.0,
13
+ "hidden_act": "silu",
14
+ "hidden_size": 2048,
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 8192,
17
+ "logits_scaling": 8.0,
18
+ "max_position_embeddings": 131072,
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": "ibm-granite/granite-3.1-2b-instruct",
26
+ "checkpoint_revision": "bbc2aed595bd38bd770263dc3ab831db9794441d",
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": 8192,
34
+ "max_topk": 256,
35
+ "n_active_tokens": 8192,
36
+ "neuronxcc_version": "2.21.33363.0+82129205",
37
+ "on_device_sampling": false,
38
+ "optimum_neuron_version": "0.4.4.dev1",
39
+ "output_logits": false,
40
+ "pp_degree": 1,
41
+ "sequence_length": 8192,
42
+ "sequence_parallel_enabled": true,
43
+ "speculation_length": 0,
44
+ "start_rank_id": 0,
45
+ "target": "trn1",
46
+ "torch_dtype": "bfloat16",
47
+ "tp_degree": 2
48
+ },
49
+ "num_attention_heads": 32,
50
+ "num_hidden_layers": 40,
51
+ "num_key_value_heads": 8,
52
+ "residual_multiplier": 0.22,
53
+ "rms_norm_eps": 1e-05,
54
+ "rope_scaling": null,
55
+ "rope_theta": 5000000.0,
56
+ "tie_word_embeddings": true,
57
+ "use_cache": true,
58
+ "vocab_size": 49155
59
+ }
neuronxcc-2.21.33363.0+82129205/MODULE_4e5d28a9a05b8ba71697+a02c3a36/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/token_generation/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
neuronxcc-2.21.33363.0+82129205/MODULE_4e5d28a9a05b8ba71697+a02c3a36/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_4e5d28a9a05b8ba71697+a02c3a36/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d13268ddb80c8558a247eb10ceee35167e93494a986409ed2d3be1eaf821e8fe
3
+ size 1122989
neuronxcc-2.21.33363.0+82129205/MODULE_4e5d28a9a05b8ba71697+a02c3a36/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:715cb890b27c64d579949ee98a97f23fd738b1ce702f2ccbc14574ab89fbf2ed
3
+ size 4957184
neuronxcc-2.21.33363.0+82129205/MODULE_4e5d28a9a05b8ba71697+a02c3a36/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4432d7a93e4a1e88d48dea0334f3c9a14a1e43c5b9b1f350c29e74aae468fa7a
3
+ size 5141125
neuronxcc-2.21.33363.0+82129205/MODULE_8e336cee3fb428dcae05+24129607/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/context_encoding/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.33363.0+82129205/MODULE_8e336cee3fb428dcae05+24129607/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_8e336cee3fb428dcae05+24129607/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28fc616c12ece846ce812a658e80999ab7dd59879008c33451a9ec4687191e9b
3
+ size 1070310
neuronxcc-2.21.33363.0+82129205/MODULE_8e336cee3fb428dcae05+24129607/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d640d44f59330bfabb52df0ae6f957016736c60e49f4362b08eb486ac98fccdf
3
+ size 15852544