danielhanchen
commited on
Commit
•
3541d82
1
Parent(s):
fe22006
Upload config
Browse files- config.json +3 -1
config.json
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "
|
3 |
"architectures": [
|
4 |
"MistralForCausalLM"
|
5 |
],
|
@@ -16,6 +16,7 @@
|
|
16 |
"num_attention_heads": 32,
|
17 |
"num_hidden_layers": 32,
|
18 |
"num_key_value_heads": 8,
|
|
|
19 |
"quantization_config": {
|
20 |
"_load_in_4bit": true,
|
21 |
"_load_in_8bit": false,
|
@@ -37,6 +38,7 @@
|
|
37 |
"tie_word_embeddings": false,
|
38 |
"torch_dtype": "bfloat16",
|
39 |
"transformers_version": "4.44.0",
|
|
|
40 |
"use_cache": true,
|
41 |
"vocab_size": 32000
|
42 |
}
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "unsloth/mistral-7b",
|
3 |
"architectures": [
|
4 |
"MistralForCausalLM"
|
5 |
],
|
|
|
16 |
"num_attention_heads": 32,
|
17 |
"num_hidden_layers": 32,
|
18 |
"num_key_value_heads": 8,
|
19 |
+
"pad_token_id": 0,
|
20 |
"quantization_config": {
|
21 |
"_load_in_4bit": true,
|
22 |
"_load_in_8bit": false,
|
|
|
38 |
"tie_word_embeddings": false,
|
39 |
"torch_dtype": "bfloat16",
|
40 |
"transformers_version": "4.44.0",
|
41 |
+
"unsloth_version": "2024.8",
|
42 |
"use_cache": true,
|
43 |
"vocab_size": 32000
|
44 |
}
|