Spaces:
Running
Running
Upload folder using huggingface_hub
Browse files
README.md
CHANGED
@@ -1,12 +1,12 @@
|
|
|
|
1 |
---
|
2 |
-
title:
|
3 |
-
emoji:
|
4 |
-
colorFrom:
|
5 |
-
colorTo:
|
6 |
sdk: gradio
|
7 |
sdk_version: 5.7.0
|
8 |
-
app_file:
|
9 |
pinned: false
|
|
|
10 |
---
|
11 |
-
|
12 |
-
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
|
|
1 |
+
|
2 |
---
|
3 |
+
title: llm_hf_transformers_main
|
4 |
+
emoji: 🔥
|
5 |
+
colorFrom: indigo
|
6 |
+
colorTo: indigo
|
7 |
sdk: gradio
|
8 |
sdk_version: 5.7.0
|
9 |
+
app_file: run.py
|
10 |
pinned: false
|
11 |
+
hf_oauth: true
|
12 |
---
|
|
|
|
requirements.txt
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
gradio-client @ git+https://github.com/gradio-app/gradio@afd75def9e979d9b255a6d22d33a3aee63b6f225#subdirectory=client/python
|
2 |
+
https://gradio-pypi-previews.s3.amazonaws.com/afd75def9e979d9b255a6d22d33a3aee63b6f225/gradio-5.7.0-py3-none-any.whl
|
3 |
+
transformers>=4.46.0
|
4 |
+
torch>=2.3.1
|
run.ipynb
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"cells": [{"cell_type": "markdown", "id": "302934307671667531413257853548643485645", "metadata": {}, "source": ["# Gradio Demo: llm_hf_transformers"]}, {"cell_type": "code", "execution_count": null, "id": "272996653310673477252411125948039410165", "metadata": {}, "outputs": [], "source": ["!pip install -q gradio transformers>=4.46.0 torch>=2.3.1 "]}, {"cell_type": "code", "execution_count": null, "id": "288918539441861185822528903084949547379", "metadata": {}, "outputs": [], "source": ["from transformers import AutoModelForCausalLM, AutoTokenizer\n", "import gradio as gr\n", "\n", "checkpoint = \"HuggingFaceTB/SmolLM2-135M-Instruct\"\n", "device = \"cpu\" # \"cuda\" or \"cpu\"\n", "tokenizer = AutoTokenizer.from_pretrained(checkpoint)\n", "model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device)\n", "\n", "def predict(message, history):\n", " history.append({\"role\": \"user\", \"content\": message})\n", " input_text = tokenizer.apply_chat_template(history, tokenize=False)\n", " inputs = tokenizer.encode(input_text, return_tensors=\"pt\").to(device) # type: ignore\n", " outputs = model.generate(inputs, max_new_tokens=100, temperature=0.2, top_p=0.9, do_sample=True)\n", " decoded = tokenizer.decode(outputs[0])\n", " response = decoded.split(\"<|im_start|>assistant\\n\")[-1].split(\"<|im_end|>\")[0]\n", " return response\n", "\n", "demo = gr.ChatInterface(predict, type=\"messages\")\n", "\n", "if __name__ == \"__main__\":\n", " demo.launch()\n"]}], "metadata": {}, "nbformat": 4, "nbformat_minor": 5}
|
run.py
ADDED
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
2 |
+
import gradio as gr
|
3 |
+
|
4 |
+
checkpoint = "HuggingFaceTB/SmolLM2-135M-Instruct"
|
5 |
+
device = "cpu" # "cuda" or "cpu"
|
6 |
+
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
|
7 |
+
model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device)
|
8 |
+
|
9 |
+
def predict(message, history):
|
10 |
+
history.append({"role": "user", "content": message})
|
11 |
+
input_text = tokenizer.apply_chat_template(history, tokenize=False)
|
12 |
+
inputs = tokenizer.encode(input_text, return_tensors="pt").to(device) # type: ignore
|
13 |
+
outputs = model.generate(inputs, max_new_tokens=100, temperature=0.2, top_p=0.9, do_sample=True)
|
14 |
+
decoded = tokenizer.decode(outputs[0])
|
15 |
+
response = decoded.split("<|im_start|>assistant\n")[-1].split("<|im_end|>")[0]
|
16 |
+
return response
|
17 |
+
|
18 |
+
demo = gr.ChatInterface(predict, type="messages")
|
19 |
+
|
20 |
+
if __name__ == "__main__":
|
21 |
+
demo.launch()
|