vincent-pro-ai/french-llm-from-scratch-GGUF

Fix FFN dimensions: gate/up_proj now [1024,4096] (F16, 702MB)

Fix FFN dimensions: gate/up_proj now [1024,4096] (Q4_K_M, 218MB)

Fix missing output_norm.weight + lm_head (F16, 558MB)

Fix missing output_norm.weight + lm_head (Q4_K_M, 177MB)

Delete french-llm-v3-mistral-f16.gguf with huggingface_hub

Delete french-llm-v3-mistral-Q4_K_M.gguf with huggingface_hub

Update README with GGUF download instructions for LM Studio

Add F16 GGUF (519MB, high precision)

Add Q4_K_M quantized GGUF (160MB, optimal for LM Studio)

Add README with usage and LM Studio steps

Add V3 mistral HF export (config+weights+tokenizer)

Upload model.safetensors

Upload french_gpt2_lm_studio.gguf

Add model_pytorch.pt

Add metadata.json

Add config.json

Add README.md

Add GGUF quantized models (Q4_K_M, Q5_K_M, Q8_0) + FP32

Add GGUF quantized models (Q4_K_M, Q5_K_M, Q8_0) + FP32

Add GGUF quantized models (Q4_K_M, Q5_K_M, Q8_0) + FP32

Add GGUF quantized models (Q4_K_M, Q5_K_M, Q8_0) + FP32

initial commit