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/d30ba84b868de5308b7c.json +58 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_44a58ce58f233ce91a8b+a02c3a36/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_44a58ce58f233ce91a8b+a02c3a36/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_44a58ce58f233ce91a8b+a02c3a36/model.hlo_module.pb +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_44a58ce58f233ce91a8b+a02c3a36/model.neff +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_44a58ce58f233ce91a8b+a02c3a36/wrapped_neff.hlo +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_fea32b55163182205276+24129607/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_fea32b55163182205276+24129607/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_fea32b55163182205276+24129607/model.hlo_module.pb +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_fea32b55163182205276+24129607/model.neff +3 -0
.gitattributes
CHANGED
|
@@ -13778,3 +13778,6 @@ neuronxcc-2.21.33363.0+82129205/MODULE_39975f855bfa447fb942+a02c3a36/wrapped_nef
|
|
| 13778 |
neuronxcc-2.21.33363.0+82129205/MODULE_543b9c5112618019f351+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13779 |
neuronxcc-2.21.33363.0+82129205/MODULE_543b9c5112618019f351+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13780 |
neuronxcc-2.21.33363.0+82129205/MODULE_790bb652e12e278e15bf+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
| 13778 |
neuronxcc-2.21.33363.0+82129205/MODULE_543b9c5112618019f351+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13779 |
neuronxcc-2.21.33363.0+82129205/MODULE_543b9c5112618019f351+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13780 |
neuronxcc-2.21.33363.0+82129205/MODULE_790bb652e12e278e15bf+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13781 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_44a58ce58f233ce91a8b+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13782 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_44a58ce58f233ce91a8b+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13783 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_fea32b55163182205276+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/d30ba84b868de5308b7c.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": 8,
|
| 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": 2,
|
| 32 |
+
"max_batch_size": 8,
|
| 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": 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": 10000000.0,
|
| 55 |
+
"tie_word_embeddings": true,
|
| 56 |
+
"use_cache": true,
|
| 57 |
+
"vocab_size": 49155
|
| 58 |
+
}
|
neuronxcc-2.21.33363.0+82129205/MODULE_44a58ce58f233ce91a8b+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_44a58ce58f233ce91a8b+a02c3a36/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_44a58ce58f233ce91a8b+a02c3a36/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7694a1e248f1bff7b695a65c157eab852f6ef5b8399fc95f474e6bc158f9b5f8
|
| 3 |
+
size 1078920
|
neuronxcc-2.21.33363.0+82129205/MODULE_44a58ce58f233ce91a8b+a02c3a36/model.neff
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eac44bc1c1eb687b4eda63bc36488d0acd5ba873968d274bcd20570fb3139599
|
| 3 |
+
size 13650944
|
neuronxcc-2.21.33363.0+82129205/MODULE_44a58ce58f233ce91a8b+a02c3a36/wrapped_neff.hlo
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f68ed49caa45dc6c99de72efb7181df547ab007f48bf16e64d0617164d1117f8
|
| 3 |
+
size 13835546
|
neuronxcc-2.21.33363.0+82129205/MODULE_fea32b55163182205276+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_fea32b55163182205276+24129607/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_fea32b55163182205276+24129607/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f27a6d30f7fa0c6ccdb3cc08e277f50e2a355ea596c110e484440658bb2b4d26
|
| 3 |
+
size 1055558
|
neuronxcc-2.21.33363.0+82129205/MODULE_fea32b55163182205276+24129607/model.neff
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:699d4e1ffa449be4ca51ee93af93c900d675c8d5b61e3b7c43dfe780f5953d3d
|
| 3 |
+
size 6267904
|