remove Phi-4 and DeepSeek Lite, change model key for ggufs 31d5efd verified hmrizal commited on 24 days ago
fix Phi-4 and DeepSeek Lite Chat by limiting max new tokens and max memory and optimizing pipeline creation 872a597 verified hmrizal commited on 25 days ago
add reset_model_cache to prevent memory leak, force cpu_only and disable 8-bit quant for Phi-4 3dda2b6 verified hmrizal commited on 26 days ago
move get_fallback_model and fallback_model to the beginning 516ac46 verified hmrizal commited on 26 days ago
update initialize_model_once and create_llm_pipeline to handle error Failed to create pipeline: 'model' 399135b verified hmrizal commited on 26 days ago
error handling model phi-4, deepseek lite, flan t5, adding fallback model 22b2e5f verified hmrizal commited on 26 days ago
add performance_tracker, pure llm w/o pandas bypass, full english 6b32859 verified hmrizal commited on May 4
modify create_fallback_pipeline and initialize_model_once without CUDA 0cca13a verified hmrizal commited on Apr 28
add model_info and replace getattr in create_llm_pipeline 67219f4 verified hmrizal commited on Apr 28
update initialize_model_once and create_llm_pipeline for GGUF model, add llama_cpp, add fallback hierarchy system 9e0e548 verified hmrizal commited on Apr 28
update initialize_model_once, create_llm_pipeline, force cpu only 8036e11 verified hmrizal commited on Apr 9
update gr.chatbot type=messages, commenting change_model_button 155810d verified hmrizal commited on Apr 9
update tambah reset_session, tidak bisa ganti model di tengah chat 77f6935 verified hmrizal commited on Apr 9
update change_model, process_file, create_llm_pipeline, explicit button to change model 7fb09f2 verified hmrizal commited on Apr 9
