Fix: use dtype= like official example for GLM model 3e8f1b5
Desorden1337 commited on
Commit History
🔥 Qwen2.5-72B - SUPPORTED LARGE MODEL 9e71a0b
Desorden1337 commited on
Fix: auto-install peft at runtime 9728fa3
Desorden1337 commited on
🔥 GLM-4.7-Flash-abliterated (31B) + 4-bit + LoRA - official loading 1d911cc
Desorden1337 commited on
🔥 Mixtral-8x7B (47B MoE) + LoRA + 4-bit ef82471
Desorden1337 commited on
Fix: use transformers from git for glm4_moe_lite support 6e9a66b
Desorden1337 commited on
Fix: Gradio every param not supported 3f15b4c
Desorden1337 commited on
Fix: remove flash-attn, optimize for L40S x4 dd4f5ac
Desorden1337 commited on
🚀 L40S x4 optimized: Flash Attention, 3 epochs, batch 64 f779211
Desorden1337 commited on
Enable A10G GPU hardware 5b73883
Desorden1337 commited on
Upload app.py with huggingface_hub bf46a25 verified
Darin Leonhart commited on
Upload run_training_direct.py with huggingface_hub 185bca3 verified
Darin Leonhart commited on
Upload app.py with huggingface_hub 95b6e6d verified
Darin Leonhart commited on
Upload app.py with huggingface_hub d03c67a verified
Darin Leonhart commited on
Upload finetune_setup.py with huggingface_hub 486fc66 verified
Darin Leonhart commited on
Upload d1337_cipher_complete.jsonl with huggingface_hub 8bb3088 verified
Darin Leonhart commited on
Upload app.py with huggingface_hub 6dc86df verified
Darin Leonhart commited on
Upload folder using huggingface_hub bfc24c4 verified
Darin Leonhart commited on
Upload folder using huggingface_hub f3d7cc1 verified
Darin Leonhart commited on
initial commit 9f1ed4b verified
Darin Leonhart commited on