Synchronizing local compiler cache.
Browse files- neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev2/inference/granite/ibm-granite/granite-3.1-2b-instruct/9e7df2d93403be65f917.json +41 -0
- neuronxcc-2.17.194.0+d312836f/MODULE_20223e04b1bfd1a5fe7e+613edded/model.neff +1 -1
- neuronxcc-2.17.194.0+d312836f/MODULE_4f54951d7f1d19705651+613edded/model.neff +1 -1
neuronxcc-2.17.194.0+d312836f/0_REGISTRY/0.2.0.dev2/inference/granite/ibm-granite/granite-3.1-2b-instruct/9e7df2d93403be65f917.json
ADDED
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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 |
+
"embedding_multiplier": 12.0,
|
12 |
+
"hidden_act": "silu",
|
13 |
+
"hidden_size": 2048,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 8192,
|
16 |
+
"logits_scaling": 8.0,
|
17 |
+
"max_position_embeddings": 131072,
|
18 |
+
"mlp_bias": false,
|
19 |
+
"model_type": "granite",
|
20 |
+
"neuron": {
|
21 |
+
"auto_cast_type": "bf16",
|
22 |
+
"batch_size": 4,
|
23 |
+
"checkpoint_id": "ibm-granite/granite-3.1-2b-instruct",
|
24 |
+
"checkpoint_revision": "374ef54e020a3ce208c65e96d6213922a87d8952",
|
25 |
+
"compiler_type": "neuronx-cc",
|
26 |
+
"compiler_version": "2.17.194.0+d312836f",
|
27 |
+
"num_cores": 2,
|
28 |
+
"sequence_length": 4096,
|
29 |
+
"task": "text-generation"
|
30 |
+
},
|
31 |
+
"num_attention_heads": 32,
|
32 |
+
"num_hidden_layers": 40,
|
33 |
+
"num_key_value_heads": 8,
|
34 |
+
"residual_multiplier": 0.22,
|
35 |
+
"rms_norm_eps": 1e-05,
|
36 |
+
"rope_scaling": null,
|
37 |
+
"rope_theta": 5000000.0,
|
38 |
+
"tie_word_embeddings": true,
|
39 |
+
"use_cache": true,
|
40 |
+
"vocab_size": 49155
|
41 |
+
}
|
neuronxcc-2.17.194.0+d312836f/MODULE_20223e04b1bfd1a5fe7e+613edded/model.neff
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 3851264
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fe296bed01ad8c356820cb7da175c0ae43f8f27a8d4a6f3562e332ebb9e1e0ef
|
3 |
size 3851264
|
neuronxcc-2.17.194.0+d312836f/MODULE_4f54951d7f1d19705651+613edded/model.neff
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 5377024
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0d7910b02278ed8bbb9bdb94247ca9cdcfe3d45a50757f5286339d48fde45cee
|
3 |
size 5377024
|