Update app.py
Browse files
app.py
CHANGED
|
@@ -10,12 +10,14 @@ model_id = "TheBloke/Yi-34B-200K-Llamafied-GPTQ"
|
|
| 10 |
gptq_config = GPTQConfig(
|
| 11 |
bits=4,
|
| 12 |
exllama_config={"version": 2}
|
|
|
|
| 13 |
)
|
| 14 |
tokenizer = YiTokenizer.from_pretrained("./")
|
| 15 |
model = AutoModelForCausalLM.from_pretrained(
|
| 16 |
model_id,
|
| 17 |
device_map="auto",
|
| 18 |
quantization_config=gptq_config
|
|
|
|
| 19 |
)
|
| 20 |
def run(message, chat_history, max_new_tokens=4056, temperature=3.5, top_p=0.9, top_k=800):
|
| 21 |
prompt = get_prompt(message, chat_history)
|
|
|
|
| 10 |
gptq_config = GPTQConfig(
|
| 11 |
bits=4,
|
| 12 |
exllama_config={"version": 2}
|
| 13 |
+
disable_exllama=True
|
| 14 |
)
|
| 15 |
tokenizer = YiTokenizer.from_pretrained("./")
|
| 16 |
model = AutoModelForCausalLM.from_pretrained(
|
| 17 |
model_id,
|
| 18 |
device_map="auto",
|
| 19 |
quantization_config=gptq_config
|
| 20 |
+
|
| 21 |
)
|
| 22 |
def run(message, chat_history, max_new_tokens=4056, temperature=3.5, top_p=0.9, top_k=800):
|
| 23 |
prompt = get_prompt(message, chat_history)
|