Upload folder using huggingface_hub
Browse files- README.md +1 -0
- config.json +1 -1
README.md
CHANGED
|
@@ -11,5 +11,6 @@ base_model:
|
|
| 11 |
|
| 12 |
# noteworthy changes
|
| 13 |
- tokenizer omits `<|audio|>` to prevent finetunes resizing the embeddings unneccessarily
|
|
|
|
| 14 |
- `generation_config.json` uses the correct EOS to end generation on end of audio + has a reasonable default temperature
|
| 15 |
- `chat_template.jinja` maps user messages to transcripts and assistant messages to audio for simple conversational context management
|
|
|
|
| 11 |
|
| 12 |
# noteworthy changes
|
| 13 |
- tokenizer omits `<|audio|>` to prevent finetunes resizing the embeddings unneccessarily
|
| 14 |
+
- `config.json` uses the correct EOS to end generation on end of audio
|
| 15 |
- `generation_config.json` uses the correct EOS to end generation on end of audio + has a reasonable default temperature
|
| 16 |
- `chat_template.jinja` maps user messages to transcripts and assistant messages to audio for simple conversational context management
|
config.json
CHANGED
|
@@ -5,7 +5,7 @@
|
|
| 5 |
"attention_bias": false,
|
| 6 |
"attention_dropout": 0.0,
|
| 7 |
"bos_token_id": 128000,
|
| 8 |
-
"eos_token_id":
|
| 9 |
"head_dim": 128,
|
| 10 |
"hidden_act": "silu",
|
| 11 |
"hidden_size": 3072,
|
|
|
|
| 5 |
"attention_bias": false,
|
| 6 |
"attention_dropout": 0.0,
|
| 7 |
"bos_token_id": 128000,
|
| 8 |
+
"eos_token_id": 128258,
|
| 9 |
"head_dim": 128,
|
| 10 |
"hidden_act": "silu",
|
| 11 |
"hidden_size": 3072,
|