Synchronizing local compiler cache.
Browse files- .gitattributes +3 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/granite/ibm-granite/granite-3.1-8b-instruct/d503755dc443f9f76c52.json +58 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_25b4636000fb60f96a63+a02c3a36/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_25b4636000fb60f96a63+a02c3a36/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_25b4636000fb60f96a63+a02c3a36/model.hlo_module.pb +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_25b4636000fb60f96a63+a02c3a36/model.neff +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_25b4636000fb60f96a63+a02c3a36/wrapped_neff.hlo +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_fcb02d64be70d882d6db+24129607/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_fcb02d64be70d882d6db+24129607/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_fcb02d64be70d882d6db+24129607/model.hlo_module.pb +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_fcb02d64be70d882d6db+24129607/model.neff +3 -0
.gitattributes
CHANGED
|
@@ -13790,3 +13790,6 @@ neuronxcc-2.21.33363.0+82129205/MODULE_da8ca89141e2db9fad2a+24129607/model.neff
|
|
| 13790 |
neuronxcc-2.21.33363.0+82129205/MODULE_2983b97a72973cfcdc67+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13791 |
neuronxcc-2.21.33363.0+82129205/MODULE_3c59002e37181a971689+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13792 |
neuronxcc-2.21.33363.0+82129205/MODULE_3c59002e37181a971689+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
| 13790 |
neuronxcc-2.21.33363.0+82129205/MODULE_2983b97a72973cfcdc67+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13791 |
neuronxcc-2.21.33363.0+82129205/MODULE_3c59002e37181a971689+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13792 |
neuronxcc-2.21.33363.0+82129205/MODULE_3c59002e37181a971689+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13793 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_25b4636000fb60f96a63+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13794 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_25b4636000fb60f96a63+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13795 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_fcb02d64be70d882d6db+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/granite/ibm-granite/granite-3.1-8b-instruct/d503755dc443f9f76c52.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "ibm-granite/granite-3.1-8b-instruct",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"GraniteForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.1,
|
| 10 |
+
"attention_multiplier": 0.0078125,
|
| 11 |
+
"dtype": "bfloat16",
|
| 12 |
+
"embedding_multiplier": 12.0,
|
| 13 |
+
"hidden_act": "silu",
|
| 14 |
+
"hidden_size": 4096,
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 12800,
|
| 17 |
+
"logits_scaling": 16.0,
|
| 18 |
+
"max_position_embeddings": 131072,
|
| 19 |
+
"mlp_bias": false,
|
| 20 |
+
"model_type": "granite",
|
| 21 |
+
"neuron": {
|
| 22 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 23 |
+
"batch_size": 32,
|
| 24 |
+
"capacity_factor": null,
|
| 25 |
+
"checkpoint_id": "ibm-granite/granite-3.1-8b-instruct",
|
| 26 |
+
"checkpoint_revision": "4009206d5fc95d2e65a7b7633e159d6e97e25d35",
|
| 27 |
+
"continuous_batching": true,
|
| 28 |
+
"ep_degree": 1,
|
| 29 |
+
"fused_qkv": true,
|
| 30 |
+
"glu_mlp": true,
|
| 31 |
+
"local_ranks_size": 8,
|
| 32 |
+
"max_batch_size": 32,
|
| 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": "0.4.3.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": 8
|
| 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": 10000000.0,
|
| 55 |
+
"tie_word_embeddings": true,
|
| 56 |
+
"use_cache": true,
|
| 57 |
+
"vocab_size": 49155
|
| 58 |
+
}
|
neuronxcc-2.21.33363.0+82129205/MODULE_25b4636000fb60f96a63+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_25b4636000fb60f96a63+a02c3a36/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_25b4636000fb60f96a63+a02c3a36/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7c52e629c4487df7861f48a15ca691617f8f0e08a52617fac13833c27893196b
|
| 3 |
+
size 1021013
|
neuronxcc-2.21.33363.0+82129205/MODULE_25b4636000fb60f96a63+a02c3a36/model.neff
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:656193fe1c2a13400bee66f3c5bd66b8520f74764ad87491b74541dc47714960
|
| 3 |
+
size 8438784
|
neuronxcc-2.21.33363.0+82129205/MODULE_25b4636000fb60f96a63+a02c3a36/wrapped_neff.hlo
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:740d47c8e9a6bc81e589cab3cad51617dbd4ebdb7d56640a2cb05f0f140ee97f
|
| 3 |
+
size 8622706
|
neuronxcc-2.21.33363.0+82129205/MODULE_fcb02d64be70d882d6db+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_fcb02d64be70d882d6db+24129607/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_fcb02d64be70d882d6db+24129607/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5e23a6da73d73ee967ae71a610ed3b77d9397b4ebfd6911964ae2fc16987ddb2
|
| 3 |
+
size 1001284
|
neuronxcc-2.21.33363.0+82129205/MODULE_fcb02d64be70d882d6db+24129607/model.neff
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:228884388d9533875e49459489aa73f5f143befbca9ac2e5fad748ba154897b7
|
| 3 |
+
size 2089984
|