Update src/streamlit_app.py
Browse files- src/streamlit_app.py +2 -30
src/streamlit_app.py
CHANGED
|
@@ -1,31 +1,3 @@
|
|
| 1 |
-
import
|
| 2 |
-
import torch
|
| 3 |
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
pipeline = transformers.pipeline(
|
| 7 |
-
"text-generation",
|
| 8 |
-
model=model_id,
|
| 9 |
-
model_kwargs={"torch_dtype": torch.bfloat16},
|
| 10 |
-
device_map="auto",
|
| 11 |
-
)
|
| 12 |
-
|
| 13 |
-
messages = [
|
| 14 |
-
{"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
|
| 15 |
-
{"role": "user", "content": "Who are you?"},
|
| 16 |
-
]
|
| 17 |
-
|
| 18 |
-
terminators = [
|
| 19 |
-
pipeline.tokenizer.eos_token_id,
|
| 20 |
-
pipeline.tokenizer.convert_tokens_to_ids("<|eot_id|>")
|
| 21 |
-
]
|
| 22 |
-
|
| 23 |
-
outputs = pipeline(
|
| 24 |
-
messages,
|
| 25 |
-
max_new_tokens=256,
|
| 26 |
-
eos_token_id=terminators,
|
| 27 |
-
do_sample=True,
|
| 28 |
-
temperature=0.6,
|
| 29 |
-
top_p=0.9,
|
| 30 |
-
)
|
| 31 |
-
print(outputs[0]["generated_text"][-1])
|
|
|
|
| 1 |
+
import gradio as gr
|
|
|
|
| 2 |
|
| 3 |
+
gr.load("models/meta-llama/Meta-Llama-3-8B-Instruct").launch()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|