Update model.yml
Browse files
model.yml
CHANGED
|
@@ -6,6 +6,8 @@ files:
|
|
| 6 |
- llama_model_path: model.gguf
|
| 7 |
|
| 8 |
# Results Preferences
|
|
|
|
|
|
|
| 9 |
top_p: 0.95
|
| 10 |
temperature: 0.7
|
| 11 |
frequency_penalty: 0
|
|
@@ -14,7 +16,6 @@ max_tokens: 4096 # Infer from base config.json -> max_position_embeddings
|
|
| 14 |
stream: true # true | false
|
| 15 |
|
| 16 |
# Engine / Model Settings
|
| 17 |
-
ngl: 33 # Infer from base config.json -> num_attention_heads
|
| 18 |
ctx_len: 4096 # Infer from base config.json -> max_position_embeddings
|
| 19 |
engine: cortex.llamacpp
|
| 20 |
prompt_template: "{system_message} [INST] {prompt} [/INST]"
|
|
|
|
| 6 |
- llama_model_path: model.gguf
|
| 7 |
|
| 8 |
# Results Preferences
|
| 9 |
+
stop:
|
| 10 |
+
- </s>
|
| 11 |
top_p: 0.95
|
| 12 |
temperature: 0.7
|
| 13 |
frequency_penalty: 0
|
|
|
|
| 16 |
stream: true # true | false
|
| 17 |
|
| 18 |
# Engine / Model Settings
|
|
|
|
| 19 |
ctx_len: 4096 # Infer from base config.json -> max_position_embeddings
|
| 20 |
engine: cortex.llamacpp
|
| 21 |
prompt_template: "{system_message} [INST] {prompt} [/INST]"
|