Upload convert_to_gguf_q8km.py with huggingface_hub

convert_to_gguf_q8km.py

@@ -0,0 +1,236 @@
+#!/usr/bin/env python3
+# /// script
+# dependencies = [
+#     "transformers>=4.36.0",
+#     "peft>=0.7.0",
+#     "torch>=2.0.0",
+#     "accelerate>=0.24.0",
+#     "huggingface_hub>=0.20.0",
+#     "sentencepiece>=0.1.99",
+#     "protobuf>=3.20.0",
+#     "numpy",
+#     "gguf",
+# ]
+# ///
+
+"""
+GGUF Conversion - Q8_K_M Quantization
+Converts LoRA fine-tuned model to GGUF with Q8_K_M quantization.
+"""
+
+import os
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from peft import PeftModel
+from huggingface_hub import HfApi
+import subprocess
+
+print("🔄 GGUF Conversion Script - Q8_K_M")
+print("=" * 60)
+
+# Configuration
+ADAPTER_MODEL = "chaddy81/qwen3-0.6b-multicode-sft"
+BASE_MODEL = "Qwen/Qwen3-0.6B"
+OUTPUT_REPO = "chaddy81/qwen3-0.6b-multicode-sft-gguf"
+username = "chaddy81"
+
+print(f"\n📦 Configuration:")
+print(f"   Base model: {BASE_MODEL}")
+print(f"   Adapter model: {ADAPTER_MODEL}")
+print(f"   Output repo: {OUTPUT_REPO}")
+print(f"   Quantization: Q8_K_M")
+
+# Step 1: Load base model and adapter
+print("\n🔧 Step 1: Loading base model and LoRA adapter...")
+base_model = AutoModelForCausalLM.from_pretrained(
+    BASE_MODEL,
+    torch_dtype=torch.float16,
+    device_map="auto",
+    trust_remote_code=True,
+)
+print("   ✅ Base model loaded")
+
+model = PeftModel.from_pretrained(base_model, ADAPTER_MODEL)
+print("   ✅ Adapter loaded")
+
+merged_model = model.merge_and_unload()
+print("   ✅ Models merged!")
+
+tokenizer = AutoTokenizer.from_pretrained(ADAPTER_MODEL, trust_remote_code=True)
+print("   ✅ Tokenizer loaded")
+
+# Step 2: Save merged model
+print("\n💾 Step 2: Saving merged model...")
+merged_dir = "/tmp/merged_model"
+merged_model.save_pretrained(merged_dir, safe_serialization=True)
+tokenizer.save_pretrained(merged_dir)
+print(f"   ✅ Merged model saved to {merged_dir}")
+
+# Step 3: Install llama.cpp
+print("\n📥 Step 3: Setting up llama.cpp...")
+
+print("   Installing build tools...")
+subprocess.run(["apt-get", "update", "-qq"], check=True, capture_output=True)
+subprocess.run(["apt-get", "install", "-y", "-qq", "build-essential", "cmake"], check=True, capture_output=True)
+print("   ✅ Build tools installed")
+
+print("   Cloning llama.cpp repository...")
+subprocess.run(["git", "clone", "--depth", "1", "https://github.com/ggerganov/llama.cpp.git", "/tmp/llama.cpp"], check=True, capture_output=True)
+print("   ✅ llama.cpp cloned")
+
+print("   Installing Python dependencies...")
+subprocess.run(["pip", "install", "-r", "/tmp/llama.cpp/requirements.txt"], check=True, capture_output=True)
+subprocess.run(["pip", "install", "sentencepiece", "protobuf"], check=True, capture_output=True)
+print("   ✅ Dependencies installed")
+
+# Step 4: Convert to GGUF (FP16)
+print("\n🔄 Step 4: Converting to GGUF format (FP16)...")
+gguf_output_dir = "/tmp/gguf_output"
+os.makedirs(gguf_output_dir, exist_ok=True)
+
+convert_script = "/tmp/llama.cpp/convert_hf_to_gguf.py"
+model_name = "qwen3-0.6b-multicode-sft"
+gguf_file = f"{gguf_output_dir}/{model_name}-f16.gguf"
+
+try:
+    result = subprocess.run(
+        ["python", convert_script, merged_dir, "--outfile", gguf_file, "--outtype", "f16"],
+        check=True, capture_output=True, text=True
+    )
+    print(result.stdout)
+except subprocess.CalledProcessError as e:
+    print(f"❌ Conversion failed!")
+    print("STDOUT:", e.stdout)
+    print("STDERR:", e.stderr)
+    raise
+print(f"   ✅ FP16 GGUF created: {gguf_file}")
+
+# Step 5: Build quantize tool and create Q8_K_M
+print("\n⚙️  Step 5: Creating Q8_K_M quantization...")
+
+print("   Building quantize tool with CMake...")
+os.makedirs("/tmp/llama.cpp/build", exist_ok=True)
+subprocess.run(
+    ["cmake", "-B", "/tmp/llama.cpp/build", "-S", "/tmp/llama.cpp", "-DGGML_CUDA=OFF"],
+    check=True, capture_output=True, text=True
+)
+subprocess.run(
+    ["cmake", "--build", "/tmp/llama.cpp/build", "--target", "llama-quantize", "-j", "4"],
+    check=True, capture_output=True, text=True
+)
+print("   ✅ Quantize tool built")
+
+quantize_bin = "/tmp/llama.cpp/build/bin/llama-quantize"
+
+# Create Q8_K_M quantization
+quant_file = f"{gguf_output_dir}/{model_name}-q8_k_m.gguf"
+print(f"   Creating Q8_K_M quantization...")
+subprocess.run([quantize_bin, gguf_file, quant_file, "Q8_K_M"], check=True, capture_output=True)
+size_mb = os.path.getsize(quant_file) / (1024 * 1024)
+print(f"   ✅ Q8_K_M: {size_mb:.1f} MB")
+
+# Step 6: Upload to Hub
+print("\n☁️  Step 6: Uploading to Hugging Face Hub...")
+api = HfApi()
+
+print(f"   Creating repository: {OUTPUT_REPO}")
+try:
+    api.create_repo(repo_id=OUTPUT_REPO, repo_type="model", exist_ok=True)
+    print("   ✅ Repository created")
+except Exception as e:
+    print(f"   ℹ️  Repository may already exist: {e}")
+
+# Upload Q8_K_M version
+print("   Uploading Q8_K_M GGUF...")
+api.upload_file(
+    path_or_fileobj=quant_file,
+    path_in_repo=f"{model_name}-q8_k_m.gguf",
+    repo_id=OUTPUT_REPO,
+)
+print("   ✅ Q8_K_M uploaded")
+
+# Create README
+print("\n📝 Creating README...")
+readme_content = f"""---
+base_model: {BASE_MODEL}
+tags:
+- gguf
+- llama.cpp
+- quantized
+- trl
+- sft
+- qwen3
+- code
+---
+
+# {model_name} GGUF
+
+GGUF conversion of [{ADAPTER_MODEL}](https://huggingface.co/{ADAPTER_MODEL}), a LoRA fine-tuned version of [{BASE_MODEL}](https://huggingface.co/{BASE_MODEL}).
+
+## Model Details
+
+- **Base Model:** {BASE_MODEL}
+- **Fine-tuned Model:** {ADAPTER_MODEL}
+- **Training:** SFT on multi-language code datasets (Codeforces, Golang, Vue/Nuxt, React)
+- **Format:** GGUF Q8_K_M quantization
+
+## Available Files
+
+| File | Quant | Description |
+|------|-------|-------------|
+| {model_name}-q8_k_m.gguf | Q8_K_M | 8-bit K-quant medium - high quality |
+
+## Usage
+
+### With llama.cpp
+
+```bash
+huggingface-cli download {OUTPUT_REPO} {model_name}-q8_k_m.gguf
+./llama-cli -m {model_name}-q8_k_m.gguf -p "Write a React component"
+```
+
+### With Ollama
+
+```bash
+# Download
+huggingface-cli download {OUTPUT_REPO} {model_name}-q8_k_m.gguf
+
+# Create Modelfile
+echo "FROM ./{model_name}-q8_k_m.gguf" > Modelfile
+
+# Create and run
+ollama create qwen3-multicode -f Modelfile
+ollama run qwen3-multicode
+```
+
+### With LM Studio
+
+1. Download the `.gguf` file
+2. Import into LM Studio
+3. Start chatting!
+
+## Training Data
+
+This model was fine-tuned on:
+- open-r1/codeforces-cots (competitive programming)
+- smcleod/golang-coder (Go)
+- kevind13/vuejs-nuxt-tailwind-codellama (Vue/Nuxt)
+- cfahlgren1/react-code-instructions (React)
+
+---
+*Converted to GGUF with Q8_K_M quantization using llama.cpp*
+"""
+
+api.upload_file(
+    path_or_fileobj=readme_content.encode(),
+    path_in_repo="README.md",
+    repo_id=OUTPUT_REPO,
+)
+print("   ✅ README uploaded")
+
+print("\n" + "=" * 60)
+print("✅ GGUF Conversion Complete!")
+print(f"📦 Repository: https://huggingface.co/{OUTPUT_REPO}")
+print(f"\n📥 Download:")
+print(f"   huggingface-cli download {OUTPUT_REPO} {model_name}-q8_k_m.gguf")
+print("=" * 60)