Upload 3 files
Browse files- config.json +30 -0
- generation_config.json +5 -0
- merges.txt +0 -0
config.json
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "./falcon-rw-1b-chat/epoch-1",
|
3 |
+
"alibi": true,
|
4 |
+
"apply_residual_connection_post_layernorm": false,
|
5 |
+
"architectures": [
|
6 |
+
"FalconForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"bias": true,
|
10 |
+
"bos_token_id": 11,
|
11 |
+
"eos_token_id": 11,
|
12 |
+
"hidden_dropout": 0.0,
|
13 |
+
"hidden_size": 2048,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"layer_norm_epsilon": 1e-05,
|
16 |
+
"max_position_embeddings": 2048,
|
17 |
+
"model_type": "falcon",
|
18 |
+
"multi_query": false,
|
19 |
+
"new_decoder_architecture": false,
|
20 |
+
"num_attention_heads": 32,
|
21 |
+
"num_hidden_layers": 24,
|
22 |
+
"num_kv_heads": 32,
|
23 |
+
"parallel_attn": false,
|
24 |
+
"rope_scaling": null,
|
25 |
+
"rope_theta": 10000.0,
|
26 |
+
"torch_dtype": "bfloat16",
|
27 |
+
"transformers_version": "4.35.2",
|
28 |
+
"use_cache": false,
|
29 |
+
"vocab_size": 50304
|
30 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token_id": 50256,
|
3 |
+
"eos_token_id": 50256,
|
4 |
+
"transformers_version": "4.35.2"
|
5 |
+
}
|
merges.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|