Commit
•
b9f2b42
1
Parent(s):
217f310
Upload Qwen2ForCausalLM
Browse files- config.json +3 -1
- generation_config.json +2 -1
config.json
CHANGED
|
@@ -5,7 +5,7 @@
|
|
| 5 |
],
|
| 6 |
"attention_dropout": 0.0,
|
| 7 |
"bos_token_id": 151643,
|
| 8 |
-
"eos_token_id":
|
| 9 |
"hidden_act": "silu",
|
| 10 |
"hidden_size": 1536,
|
| 11 |
"initializer_range": 0.02,
|
|
@@ -16,12 +16,14 @@
|
|
| 16 |
"num_attention_heads": 12,
|
| 17 |
"num_hidden_layers": 28,
|
| 18 |
"num_key_value_heads": 2,
|
|
|
|
| 19 |
"rms_norm_eps": 1e-06,
|
| 20 |
"rope_theta": 1000000.0,
|
| 21 |
"sliding_window": null,
|
| 22 |
"tie_word_embeddings": true,
|
| 23 |
"torch_dtype": "bfloat16",
|
| 24 |
"transformers_version": "4.44.2",
|
|
|
|
| 25 |
"use_cache": true,
|
| 26 |
"use_sliding_window": false,
|
| 27 |
"vocab_size": 151936
|
|
|
|
| 5 |
],
|
| 6 |
"attention_dropout": 0.0,
|
| 7 |
"bos_token_id": 151643,
|
| 8 |
+
"eos_token_id": 151645,
|
| 9 |
"hidden_act": "silu",
|
| 10 |
"hidden_size": 1536,
|
| 11 |
"initializer_range": 0.02,
|
|
|
|
| 16 |
"num_attention_heads": 12,
|
| 17 |
"num_hidden_layers": 28,
|
| 18 |
"num_key_value_heads": 2,
|
| 19 |
+
"pad_token_id": 151665,
|
| 20 |
"rms_norm_eps": 1e-06,
|
| 21 |
"rope_theta": 1000000.0,
|
| 22 |
"sliding_window": null,
|
| 23 |
"tie_word_embeddings": true,
|
| 24 |
"torch_dtype": "bfloat16",
|
| 25 |
"transformers_version": "4.44.2",
|
| 26 |
+
"unsloth_fixed": true,
|
| 27 |
"use_cache": true,
|
| 28 |
"use_sliding_window": false,
|
| 29 |
"vocab_size": 151936
|
generation_config.json
CHANGED
|
@@ -5,7 +5,8 @@
|
|
| 5 |
151645,
|
| 6 |
151643
|
| 7 |
],
|
| 8 |
-
"
|
|
|
|
| 9 |
"repetition_penalty": 1.1,
|
| 10 |
"temperature": 0.7,
|
| 11 |
"top_k": 20,
|
|
|
|
| 5 |
151645,
|
| 6 |
151643
|
| 7 |
],
|
| 8 |
+
"max_length": 32768,
|
| 9 |
+
"pad_token_id": 151665,
|
| 10 |
"repetition_penalty": 1.1,
|
| 11 |
"temperature": 0.7,
|
| 12 |
"top_k": 20,
|