Upload LlamaForCausalLM
Browse files- config.json +103 -0
- generation_config.json +6 -0
- model.safetensors +3 -0
config.json
ADDED
|
@@ -0,0 +1,103 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "meta-llama/Llama-2-7b-hf",
|
| 3 |
+
"a_ps": [
|
| 4 |
+
[
|
| 5 |
+
"dp",
|
| 6 |
+
"fsdp"
|
| 7 |
+
],
|
| 8 |
+
"sp",
|
| 9 |
+
"tp",
|
| 10 |
+
null
|
| 11 |
+
],
|
| 12 |
+
"architectures": [
|
| 13 |
+
"LlamaForCausalLM"
|
| 14 |
+
],
|
| 15 |
+
"attention_bias": false,
|
| 16 |
+
"attention_dropout": 0.0,
|
| 17 |
+
"axis_dims": [
|
| 18 |
+
1,
|
| 19 |
+
-1,
|
| 20 |
+
1,
|
| 21 |
+
1
|
| 22 |
+
],
|
| 23 |
+
"axis_names": [
|
| 24 |
+
"dp",
|
| 25 |
+
"fsdp",
|
| 26 |
+
"tp",
|
| 27 |
+
"sp"
|
| 28 |
+
],
|
| 29 |
+
"b_ps": [
|
| 30 |
+
[
|
| 31 |
+
"dp",
|
| 32 |
+
"fsdp"
|
| 33 |
+
],
|
| 34 |
+
null,
|
| 35 |
+
null,
|
| 36 |
+
null
|
| 37 |
+
],
|
| 38 |
+
"backend": null,
|
| 39 |
+
"bits": null,
|
| 40 |
+
"bos_token_id": 1,
|
| 41 |
+
"c_max_position_embeddings": 2048,
|
| 42 |
+
"easy_method": "train",
|
| 43 |
+
"embd_pdrop": 0.0,
|
| 44 |
+
"eos_token_id": 2,
|
| 45 |
+
"fcm_max_ratio": 0.0,
|
| 46 |
+
"fcm_min_ratio": 0.0,
|
| 47 |
+
"flash_attn_key_chunk_size": 1024,
|
| 48 |
+
"flash_attn_query_chunk_size": 1024,
|
| 49 |
+
"freq_max_position_embeddings": 2048,
|
| 50 |
+
"hidden_act": "silu",
|
| 51 |
+
"hidden_size": 2048,
|
| 52 |
+
"initializer_range": 0.02,
|
| 53 |
+
"intermediate_size": 5632,
|
| 54 |
+
"k_ps": [
|
| 55 |
+
[
|
| 56 |
+
"dp",
|
| 57 |
+
"fsdp"
|
| 58 |
+
],
|
| 59 |
+
"sp",
|
| 60 |
+
"tp",
|
| 61 |
+
null
|
| 62 |
+
],
|
| 63 |
+
"max_position_embeddings": 2048,
|
| 64 |
+
"model_type": "llama",
|
| 65 |
+
"num_attention_heads": 32,
|
| 66 |
+
"num_hidden_layers": 22,
|
| 67 |
+
"num_key_value_heads": 4,
|
| 68 |
+
"number_rep_kv": 1,
|
| 69 |
+
"pretraining_tp": 1,
|
| 70 |
+
"q_ps": [
|
| 71 |
+
[
|
| 72 |
+
"dp",
|
| 73 |
+
"fsdp"
|
| 74 |
+
],
|
| 75 |
+
"sp",
|
| 76 |
+
"tp",
|
| 77 |
+
null
|
| 78 |
+
],
|
| 79 |
+
"resid_pdrop": 0.0,
|
| 80 |
+
"rms_norm_eps": 1e-05,
|
| 81 |
+
"rope_scaling": null,
|
| 82 |
+
"rope_theta": 10000.0,
|
| 83 |
+
"scan_layers": true,
|
| 84 |
+
"scan_mlp_chunk_size": 1024,
|
| 85 |
+
"tie_word_embeddings": false,
|
| 86 |
+
"torch_dtype": "float16",
|
| 87 |
+
"transformers_version": "4.36.0",
|
| 88 |
+
"use_cache": true,
|
| 89 |
+
"use_flash_attention": false,
|
| 90 |
+
"use_pjit_attention_force": false,
|
| 91 |
+
"use_sacn_mlp": false,
|
| 92 |
+
"use_shard_map": false,
|
| 93 |
+
"v_ps": [
|
| 94 |
+
[
|
| 95 |
+
"dp",
|
| 96 |
+
"fsdp"
|
| 97 |
+
],
|
| 98 |
+
"sp",
|
| 99 |
+
"tp",
|
| 100 |
+
null
|
| 101 |
+
],
|
| 102 |
+
"vocab_size": 32000
|
| 103 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.36.0"
|
| 6 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ac659dbd72708db677184addb953840a37bb70fbd94b364461bee5de70b7b6be
|
| 3 |
+
size 2200119664
|