Upload 6 files

README.md

@@ -1,13 +1,17 @@
 ---
-title: Chat Gpt
-emoji: ⚡
+title: GPT-OSS-120B on AMD MI300X
+emoji: 💻
 colorFrom: red
-colorTo: purple
+colorTo: pink
 sdk: gradio
-sdk_version: 5.46.1
+sdk_version: 5.36.2
 app_file: app.py
 pinned: false
-short_description: chat gpt
+license: apache-2.0
+short_description: 'gpt-oss-120b on AMD MI300X GPUs'
+models:
+  - openai/gpt-oss-120b
 ---
+***DISCLAIMER:*** Analysis is provided along with final response to allow users to gain insight into model's chain of thought, but may contain content not deemed suitable to include in final response.
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,115 @@
+import os, re, logging, gradio as gr
+from openai import OpenAI
+from gateway import request_generation
+from utils import LATEX_DELIMS
+    
+openai_api_key = os.getenv("API_KEY")
+openai_api_base = os.getenv("API_ENDPOINT")
+MODEL = os.getenv("MODEL_NAME", "")
+client = OpenAI(api_key=openai_api_key, base_url=openai_api_base)
+MAX_NEW_TOKENS = int(os.getenv("MAX_NEW_TOKENS", 1024))
+CONCURRENCY_LIMIT = int(os.getenv("CONCURRENCY_LIMIT", 20))
+QUEUE_SIZE = int(os.getenv("QUEUE_SIZE", CONCURRENCY_LIMIT * 4))
+
+logging.basicConfig(level=logging.INFO)
+
+def format_analysis_response(text):
+    m = re.search(r"analysis(.*?)assistantfinal", text, re.DOTALL)
+    if m:
+        reasoning = m.group(1).strip()
+        response = text.split("assistantfinal", 1)[-1].strip()
+        return (
+            f"**🤔 Analysis:**\n\n*{reasoning}*\n\n---\n\n"
+            f"**💬 Response:**\n\n{response}"
+        )
+    return text.strip()
+
+def generate(message, history,
+             system_prompt, temperature,
+             frequency_penalty, presence_penalty,
+             max_new_tokens):
+
+    if not message.strip():
+        yield "Please enter a prompt."
+        return
+
+    msgs = []
+    for h in history:
+        if isinstance(h, dict):
+            msgs.append(h)
+        elif isinstance(h, (list, tuple)) and len(h) == 2:
+            u, a = h
+            if u: msgs.append({"role": "user", "content": u})
+            if a: msgs.append({"role": "assistant", "content": a})
+
+    logging.info(f"[User] {message}")
+    logging.info(f"[System] {system_prompt} | Temp={temperature}")
+
+    collected, buffer = "", ""
+    yielded_once = False
+
+    try:
+        for delta in request_generation(
+            api_key=openai_api_key, api_base=openai_api_base,
+            message=message, system_prompt=system_prompt,
+            model_name=MODEL, chat_history=msgs,
+            temperature=temperature,
+            frequency_penalty=frequency_penalty,
+            presence_penalty=presence_penalty,
+            max_new_tokens=max_new_tokens,
+        ):
+            if not delta:
+                continue
+
+            collected += delta
+            buffer += delta
+
+            if not yielded_once:
+                yield delta
+                buffer = ""
+                yielded_once = True
+                continue
+
+            if "\n" in buffer or len(buffer) > 150:
+                yield collected
+                buffer = ""
+
+        final = format_analysis_response(collected)
+        if final.count("$") % 2:
+            final += "$"
+        yield final
+
+    except Exception as e:
+        logging.exception("Stream failed")
+        yield f"❌ Error: {e}"
+
+chatbot_ui = gr.ChatInterface(
+    fn=generate,
+    type="messages",
+    chatbot=gr.Chatbot(
+        label="OSS vLLM Chatbot",
+        type="messages",
+        scale=2,
+        height=600,
+        latex_delimiters=LATEX_DELIMS,
+    ),
+    stop_btn=True,
+    additional_inputs=[
+        gr.Textbox(label="System prompt", value="You are a helpful assistant.", lines=2),
+        gr.Slider(label="Temperature", minimum=0.0, maximum=1.0, step=0.1, value=0.7),
+    ],
+    examples=[
+        ["Explain the difference between supervised and unsupervised learning."],
+        ["Summarize the plot of Inception in two sentences."],
+        ["Show me the LaTeX for the quadratic formula."],
+        ["What are advantages of AMD Instinct MI300X GPU?"],
+        ["Derive the gradient of softmax cross-entropy loss."],
+        ["Explain why ∂/∂x xⁿ = n·xⁿ⁻¹ holds."],
+    ],
+    # title="Open-source GPT-OSS-120B on AMD MI300X",
+    title=" GPT-OSS-120B on AMD MI300X",
+    description="This Space is an Alpha release that demonstrates gpt-oss-120b model running on AMD MI300 infrastructure. The space is built with Apache 2.0 License.",
+)
+if __name__ == "__main__":
+    chatbot_ui.queue(max_size=QUEUE_SIZE,
+                     default_concurrency_limit=CONCURRENCY_LIMIT).launch()

gateway.py

@@ -0,0 +1,69 @@
+import logging
+from openai import OpenAI
+from typing import List, Generator, Optional
+
+logging.basicConfig(level=logging.INFO)
+
+def request_generation(
+    api_key: str,
+    api_base: str,
+    message: str,
+    system_prompt: str,
+    model_name: str,
+    chat_history: Optional[List[dict]] = None,
+    temperature: float = 0.3,
+    frequency_penalty: float = 0.0,
+    presence_penalty: float = 0.0,
+    max_new_tokens: int = 1024,
+    tools: Optional[List[dict]] = None,
+    tool_choice: Optional[str] = None,
+) -> Generator[str, None, None]:
+    """
+    Sends a streaming chat request to an OpenAI-compatible backend using the official OpenAI client.
+    Buffers output to improve LaTeX rendering.
+    """
+    client = OpenAI(api_key=api_key, base_url=api_base)
+
+    messages = [{"role": "system", "content": system_prompt}]
+    if chat_history:
+        messages.extend(chat_history)
+    messages.append({"role": "user", "content": message})
+
+    request_args = {
+        "model": model_name,
+        "messages": messages,
+        "temperature": temperature,
+        "frequency_penalty": frequency_penalty,
+        "presence_penalty": presence_penalty,
+        "max_tokens": max_new_tokens,
+        "stream": True,
+    }
+
+    if tools:
+        request_args["tools"] = tools
+    if tool_choice:
+        request_args["tool_choice"] = tool_choice
+
+    logging.info(f"[Gateway] Request to {api_base} | Model: {model_name}")
+
+    try:
+        stream = client.chat.completions.create(**request_args)
+
+        collected = ""
+        buffer = ""
+
+        for chunk in stream:
+            delta = chunk.choices[0].delta.content or ""
+            collected += delta
+            buffer += delta
+
+            if "\n" in buffer or len(buffer) > 150:
+                yield buffer
+                buffer = ""
+
+        if buffer:
+            yield buffer
+
+    except Exception as e:
+        logging.exception("[Gateway] Streaming failed")
+        yield f"Error: {e}"

gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

requirements.txt

@@ -0,0 +1 @@
+openai

utils.py

@@ -0,0 +1,9 @@
+
+# ----------------------------------------------------------------------
+# KaTeX delimiter config for Gradio
+# ----------------------------------------------------------------------
+
+LATEX_DELIMS = [
+    {"left": "\\[", "right": "\\]", "display": True},  
+    {"left": "\\(", "right": "\\)", "display": False}, 
+]