Upload README.md with huggingface_hub

README.md

@@ -25,10 +25,10 @@ This is a **GGUF** quantized version of [zai-org/AutoGLM-Phone-9B-Multilingual](
 
 | File | Quantization | Size | VRAM | Description |
 |------|-------------|------|------|-------------|
-| `AutoGLM-Phone-9B-Multilingual-q4_k_m.gguf` | Q4_K_M | 5.7G | 16GB | Performance balanced |
-| `AutoGLM-Phone-9B-Multilingual-q5_k_m.gguf` | Q5_K_M | 6.6G | 20GB | High quality |
-| `AutoGLM-Phone-9B-Multilingual-q6_k.gguf` | Q6_K | 7.7G | 24GB | Excellent quality |
-| `AutoGLM-Phone-9B-Multilingual-q8_0.gguf` | Q8_0 | 9.4G | 24GB+ | Best quality |
+| `AutoGLM-Phone-9B-Multilingual-q4_k_m.gguf` | Q4_K_M | 5.7G | ~10GB | Performance balanced |
+| `AutoGLM-Phone-9B-Multilingual-q5_k_m.gguf` | Q5_K_M | 6.6G | ~11GB | High quality |
+| `AutoGLM-Phone-9B-Multilingual-q6_k.gguf` | Q6_K | 7.7G | ~12GB | Excellent quality |
+| `AutoGLM-Phone-9B-Multilingual-q8_0.gguf` | Q8_0 | 9.4G | ~14GB | Best quality |
 | `mmproj-AutoGLM-Phone-9B-Multilingual-F16.gguf` | F16 | 1.7G | - | Vision Encoder (required) |
 
 **Total storage**: ~31GB (all quantizations + vision encoder)
@@ -78,14 +78,16 @@ python main.py \
 
 ## 💻 Hardware Requirements
 
-### Quick Reference
+### Quick Reference (Tested on RTX 4090)
 
-| Quantization | Model Size | Vision Encoder | Total | Min VRAM | Recommended GPU | Quality |
-|--------------|------------|----------------|-------|----------|-----------------|---------|
-| Q4_K_M | 5.7G | 1.7G | ~7.4G | 16GB | RTX 4080 / 4070 Ti | Good |
-| Q5_K_M | 6.6G | 1.7G | ~8.3G | 20GB | RTX 4090 / 3090 | Very Good |
-| Q6_K | 7.7G | 1.7G | ~9.4G | 24GB | RTX 4090 | Excellent |
-| Q8_0 | 9.4G | 1.7G | ~11.1G | 24GB+ | RTX 4090 / A6000 | Best |
+| Quantization | Model Size | Vision Encoder | Total | Actual VRAM* | Quality |
+|--------------|------------|----------------|-------|--------------|---------|
+| Q4_K_M | 5.7G | 1.7G | ~7.4G | ~10GB | Good |
+| Q5_K_M | 6.6G | 1.7G | ~8.3G | ~11GB | Very Good |
+| Q6_K | 7.7G | 1.7G | ~9.4G | ~12GB | Excellent |
+| Q8_0 | 9.4G | 1.7G | ~11.1G | ~14GB | Best |
+
+\*VRAM usage measured with `--flash-attn on` and all layers on GPU (`-ngl 99`)
 
 ### System Requirements
 
@@ -101,6 +103,7 @@ python main.py \
 - **KV Cache**: Quantized to Q8_0 to reduce memory usage
 - **Batch Size**: Optimized for RTX 4090 (adjust based on your GPU)
 - **Context**: Supports up to 32K tokens with M-RoPE
+- **All layers on GPU**: Set `-ngl 99` to offload all transformer layers to GPU
 
 ## 🎯 Recommended Usage
 
@@ -112,15 +115,15 @@ Use **Q5_K_M** or **Q6_K** for the best balance between quality and performance:
 
 ### For Maximum Quality
 Use **Q8_0** when:
-- VRAM is not a constraint
-- You need the highest possible accuracy
-- Running complex multi-step tasks
+- You want the highest possible accuracy
+- Running on RTX 4090 or better
+- Complex multi-step GUI automation tasks
 
-### For Constrained Hardware
+### For Consumer GPUs
 Use **Q4_K_M** when:
-- Limited VRAM (16GB)
+- Limited VRAM (12GB cards like RTX 4070)
 - Need faster inference
-- Running on consumer GPUs
+- Running on gaming GPUs
 
 ## 📄 License
 
@@ -137,3 +140,4 @@ https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual
 
 **Conversion Date**: 2025-12-29
 **llama.cpp Version**: latest (with GLM4V support)
+**Tested Hardware**: RTX 4090 24GB