Commit
410a627
·
verified ·
1 Parent(s): d38ba26

Synchronizing local compiler cache.

Browse files
.gitattributes CHANGED
@@ -6092,3 +6092,6 @@ neuronxcc-2.22.12471.0+b4a00d10/MODULE_7d1031ed0a2a7c7b50f6+519f203d/model.neff
6092
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_ac19d0c7694ae5e0b3be+519f203d/model.neff filter=lfs diff=lfs merge=lfs -text
6093
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_e7679a6b37697909b0d2+84337dd9/model.neff filter=lfs diff=lfs merge=lfs -text
6094
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_e7679a6b37697909b0d2+84337dd9/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
 
 
 
 
6092
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_ac19d0c7694ae5e0b3be+519f203d/model.neff filter=lfs diff=lfs merge=lfs -text
6093
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_e7679a6b37697909b0d2+84337dd9/model.neff filter=lfs diff=lfs merge=lfs -text
6094
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_e7679a6b37697909b0d2+84337dd9/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
6095
+ neuronxcc-2.21.33363.0+82129205/MODULE_089479b62bff8cacc87f+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
6096
+ neuronxcc-2.21.33363.0+82129205/MODULE_089479b62bff8cacc87f+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
6097
+ neuronxcc-2.21.33363.0+82129205/MODULE_a0f36ad51d6ea0dcdcbb+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.4.dev2/granite/ibm-granite/granite-3.1-2b-instruct/836a0af195498bb3e459.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 4,
24
+ "capacity_factor": null,
25
+ "checkpoint_id": "ibm-granite/granite-3.1-2b-instruct",
26
+ "checkpoint_revision": "bbc2aed595bd38bd770263dc3ab831db9794441d",
27
+ "continuous_batching": true,
28
+ "ep_degree": 1,
29
+ "fused_qkv": true,
30
+ "glu_mlp": true,
31
+ "local_ranks_size": 2,
32
+ "max_batch_size": 4,
33
+ "max_context_length": 4096,
34
+ "max_topk": 256,
35
+ "n_active_tokens": 4096,
36
+ "neuronxcc_version": "2.21.33363.0+82129205",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.4.dev2",
39
+ "output_logits": false,
40
+ "pp_degree": 1,
41
+ "sequence_length": 4096,
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": 32,
49
+ "num_hidden_layers": 40,
50
+ "num_key_value_heads": 8,
51
+ "residual_multiplier": 0.22,
52
+ "rms_norm_eps": 1e-05,
53
+ "rope_scaling": null,
54
+ "rope_theta": 5000000.0,
55
+ "tie_word_embeddings": true,
56
+ "use_cache": true,
57
+ "vocab_size": 49155
58
+ }
neuronxcc-2.21.33363.0+82129205/MODULE_089479b62bff8cacc87f+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_089479b62bff8cacc87f+a02c3a36/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_089479b62bff8cacc87f+a02c3a36/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7524acc2c61ca9b6b046acf11c51438e1276c64c121b3453edba59089b340e2
3
+ size 1080658
neuronxcc-2.21.33363.0+82129205/MODULE_089479b62bff8cacc87f+a02c3a36/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f21f6bbda57bf7e070cdc973f3fa7c2644644fbf240be8541b7065349c8dcef8
3
+ size 6933504
neuronxcc-2.21.33363.0+82129205/MODULE_089479b62bff8cacc87f+a02c3a36/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:10393c2cab5611ae66184955d37e5ee955eb6f40016323dcc12ac2c8d4bcc491
3
+ size 7117697
neuronxcc-2.21.33363.0+82129205/MODULE_16c5142e54049280b3a6+24129607/model.neff CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6d47b2099e1f344db8cd4f2ae01b0eac58d96d14b340a38c9c1ae4ff07ff5939
3
  size 71302144
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68389367bb30af25de76a4c417e2d0c446b2562faa97b2cbb36aa2a9256dae0d
3
  size 71302144
neuronxcc-2.21.33363.0+82129205/MODULE_a0f36ad51d6ea0dcdcbb+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_a0f36ad51d6ea0dcdcbb+24129607/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_a0f36ad51d6ea0dcdcbb+24129607/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd08578a4e4e68ff261cc258981b4fea2a3be84d1f5c87f5287e9c016bb03871
3
+ size 1050679
neuronxcc-2.21.33363.0+82129205/MODULE_a0f36ad51d6ea0dcdcbb+24129607/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1dfad88dded44abf213ec8d96d06594c888361b6e60c3df2af7cf59eb0b3ad5d
3
+ size 5827584
neuronxcc-2.21.33363.0+82129205/MODULE_bcbfbafeff377b2f5566+a02c3a36/model.neff CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8182ac1ce77047e9988d08414cb4d501647ae5e330d05427579ff4809e17c177
3
  size 1721344
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c0f0db943bae305a68c80774a9fe26ed9f803348b9759982bd1bb945756f1787
3
  size 1721344
neuronxcc-2.21.33363.0+82129205/MODULE_bcbfbafeff377b2f5566+a02c3a36/wrapped_neff.hlo CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d647b453b48c25e6365291bcc24d360ce075c338a776de1588fefd0681f53ffb
3
  size 1877650
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c890850f31c43d590c8728eaa8097b04b52f04486c7c1b77efb67e947a11b328
3
  size 1877650