Upload it

Dockerfile

@@ -0,0 +1,24 @@
+# Use an alias for the base image for easier updates
+FROM python:3.10 as base
+
+# Set model
+ENV MODEL=seyf1elislam/WestKunai-Hermes-7b
+ENV QUANT=Q4_K_M
+ENV CHAT_TEMPLATE=chatml
+
+# Set the working directory
+WORKDIR /app
+
+# Install Python requirements
+COPY ./requirements.txt /app/
+RUN pip install --no-cache-dir --upgrade -r requirements.txt
+
+# Download model
+RUN MODEL_NAME_FILE=$(echo ${MODEL#*/} | tr '[:upper:]' '[:lower:]' | sed 's/-gguf$//') && \
+    wget https://huggingface.co/${MODEL}/resolve/main/${MODEL_NAME_FILE}.${QUANT}.gguf -O model.gguf
+
+# Copy the rest of your application
+COPY . .
+
+# Command to run the application
+CMD ["python", "app.py"]

README.md

@@ -1,6 +1,6 @@
 ---
 title: WestKunai Hermes 7b
-emoji: 🚀
+emoji: ⚔︎
 colorFrom: pink
 colorTo: yellow
 sdk: docker

app.py

@@ -0,0 +1,58 @@
+import os
+import json
+import gradio as gr
+from llama_cpp import Llama
+
+# Get environment variables
+model_id = os.getenv('MODEL')
+quant = os.getenv('QUANT')
+chat_template = os.getenv('CHAT_TEMPLATE')
+
+# Interface variables
+model_name = model_id.split('/')[1].split('-GGUF')[0]
+title = f"{model_name}"
+description = f"Chat with <a href=\"https://huggingface.co/{model_id}\">{model_name}</a> in GGUF format ({quant})!"
+
+# Initialize the LLM
+llm = Llama(model_path="model.gguf",
+            n_ctx=32768,
+            n_threads=2,
+            n_vocab=32002,
+            n_gpu_layers=-1,
+            chat_format=chat_template)
+
+# Function for streaming chat completions
+def chat_stream_completion(message, history, system_prompt):
+    messages_prompts = [{"role": "system", "content": system_prompt}]
+    for human, assistant in history:
+        messages_prompts.append({"role": "user", "content": human})
+        messages_prompts.append({"role": "assistant", "content": assistant})
+    messages_prompts.append({"role": "user", "content": message})
+
+    response = llm.create_chat_completion(
+        messages=messages_prompts,
+        stream=True
+    )
+    message_repl = ""
+    for chunk in response:
+        if len(chunk['choices'][0]["delta"]) != 0 and "content" in chunk['choices'][0]["delta"]:
+            message_repl = message_repl + chunk['choices'][0]["delta"]["content"]
+        yield message_repl
+
+# Gradio chat interface
+gr.ChatInterface(
+    fn=chat_stream_completion,
+    title=title,
+    description=description,
+    additional_inputs=[gr.Textbox("You are helpful assistant.")],
+    additional_inputs_accordion="📝 System prompt",
+    examples=[
+        ['Can you solve the equation 2x + 3 = 11 for x?'],
+        ['Write an epic poem about Ancient Rome.'],
+        ['Who was the first person to walk on the Moon?'],
+        ['Use a list comprehension to create a list of squares for numbers from 1 to 10.'],
+        ['Recommend some popular science fiction books.'],
+        ['Can you write a short story about a time-traveling detective?']
+    ],
+    theme = gr.themes.Base()
+).queue().launch(server_name="0.0.0.0")

requirements.txt

@@ -0,0 +1,2 @@
+llama-cpp-python
+gradio