forum

1495 threads · Page 12 of 30

FastVisionModel 4bit + save_pretrained_merged 5 messages
Lower accuracy after merging adapters 7 messages
How to increase VRAM allocation for unsloth ? (Or inference speed) 18 messages
Qwen3-8b CPT incomplete answer issue 4 messages
gemma3-n hackathon 133 messages
make bnb 4bit version of mistralai/Mistral-7B-v0.3 2 messages
will you guys make a finetuning notebook for falcon h1? 17 messages
Open
I really dont unterstand the problem 29 messages
Open
LLM.get_vllm_config doesn't exist 10 messages
SFTFinetune Image Gemma3n AssertionError: expected size 16==16, stride 35 messages
Persistent NaN grad norm when finetuning Gemma 3n on images 5 messages
Error while trying to Finetune a already finetuned Gemma3n model 5 messages
Open
Inférence endpoints 2 messages
save_to_gguf_generic() got multiple values for argument 'quantization_method' 49 messages
Adding a new language to a TTS model 30 messages
Open
Current Dynamic 2.0 GGUF support in vLLM? 2 messages
Gemma 3n Quantization Support for Image Inference 8 messages
Async GRPO reward functions 3 messages
Possibility of accelerating MoE training? 2 messages
Open
Do you guys offer SaaS version of it? 10 messages
Very slow Gemma 3 inference speeds 24 messages
Open
when converting fine tuned model to GGUF i got this error : 5 messages
Learning rate stays at 0 for 15 epochs when finetuning Gemma 3n 4 messages
Error: 'str' object has no attribute 'str' in ModernBERT-large 27 messages
Solved
converting the unsloth/gemma-3n-E4B-it-unsloth-bnb-4bit and learned lora adapters into tflite format 7 messages
Open
Save fine-tuned model error 4 messages
I got this error when running the fine tuning sample code for Gemma 3n : 2 messages
Merge and reload gives very different output for Gemma-4b 162 messages
How to run vLLM when GRPO Gemma-3n 26 messages
ObjMismatchError 2 messages
Open
Help with distilling/fine tuning Qwen2.5:7b 12 messages
Open
Gemma 3 Inferencing Error 45 messages
Solved
does the seq length we train with have to = the context length I want to use the model with? 3 messages
Fine-tuned Qwen2.5-7B model loops infinitely in Ollama but works fine with transformers 8 messages
Weird Token Management In llama.cpp 3 messages
config error for `unsloth/medgemma-27b-it-GGUF` 5 messages
CheckpointError While fine tuning Gemma 3n 4 messages
TRL checkpoints with LoRA not working 4 messages
Open
Training llama 3.1 8b model on odd data 10 messages
RuntimeError: expected scalar type BFloat16 but found Half 2 messages
Support for finetuning internvl 2.5 models 4 messages
Gemma3n gguf saving broken ? 3 messages
Help in training Mistral 3.2 on photos dataset 2 messages
Fine-tuning on kaggle environment 25 messages
Cannot start any Notebook 14 messages
Understanding GPU Memory Utilization with vLLMs 58 messages
Open
Help with fine-tuning llama3.1 react style agent 11 messages
Open
RuntimeError: Bad in-place call: input tensor size [2097152, 1] and output tensor size [1024, 4096] 7 messages
Open
Can't load g3mma 3n E2B locally on 4070 mobile 11 messages
Current state of Gemma3n support 6 messages