danielhanchen
commited on
Upload LlamaForCausalLM
Browse files- config.json +3 -1
- generation_config.json +3 -1
config.json
CHANGED
@@ -22,6 +22,7 @@
|
|
22 |
"num_attention_heads": 24,
|
23 |
"num_hidden_layers": 28,
|
24 |
"num_key_value_heads": 8,
|
|
|
25 |
"pretraining_tp": 1,
|
26 |
"rms_norm_eps": 1e-05,
|
27 |
"rope_scaling": {
|
@@ -34,7 +35,8 @@
|
|
34 |
"rope_theta": 500000.0,
|
35 |
"tie_word_embeddings": true,
|
36 |
"torch_dtype": "bfloat16",
|
37 |
-
"transformers_version": "4.
|
|
|
38 |
"use_cache": true,
|
39 |
"vocab_size": 128256
|
40 |
}
|
|
|
22 |
"num_attention_heads": 24,
|
23 |
"num_hidden_layers": 28,
|
24 |
"num_key_value_heads": 8,
|
25 |
+
"pad_token_id": 128004,
|
26 |
"pretraining_tp": 1,
|
27 |
"rms_norm_eps": 1e-05,
|
28 |
"rope_scaling": {
|
|
|
35 |
"rope_theta": 500000.0,
|
36 |
"tie_word_embeddings": true,
|
37 |
"torch_dtype": "bfloat16",
|
38 |
+
"transformers_version": "4.48.0",
|
39 |
+
"unsloth_fixed": true,
|
40 |
"use_cache": true,
|
41 |
"vocab_size": 128256
|
42 |
}
|
generation_config.json
CHANGED
@@ -6,7 +6,9 @@
|
|
6 |
128008,
|
7 |
128009
|
8 |
],
|
|
|
|
|
9 |
"temperature": 0.6,
|
10 |
"top_p": 0.9,
|
11 |
-
"transformers_version": "4.
|
12 |
}
|
|
|
6 |
128008,
|
7 |
128009
|
8 |
],
|
9 |
+
"max_length": 131072,
|
10 |
+
"pad_token_id": 128004,
|
11 |
"temperature": 0.6,
|
12 |
"top_p": 0.9,
|
13 |
+
"transformers_version": "4.48.0"
|
14 |
}
|