forum

1495 threads · Page 18 of 30

Being forced to log into wandb.ai? 2 messages
Solved
Issues while merging base model(llama3.2) with adapter(lora) 3 messages
What does it mean when a model seems like it doesnt know anything about your dataset? 11 messages
Open
ValueError: numpy.dtype size changed, may indicate binary incompatibility. Expected 96 from C header 21 messages
Solved
Fill in the middle fine tuning 11 messages
Solved
Is learned prompt tuning available in Unsloth? 3 messages
Multi-Turn GRPO Brainstorming: Non-deterministic tool feedback 8 messages
Fine tuning Llama-3-8B-Lexi-Uncensored on custom dataset - LR not declining 4 messages
Fine Tuning for Regression 4 messages
Gemma 3 notebook: saving GGUF broken 2 messages
Open
RuntimeError: CUDA device-side assert triggered (Unsloth GRPO, Multi-GPU, Qwen 2.5 14B Training) 2 messages
Open
How do I get a ref to a model/lora I created? 4 messages
4 bit load error 4 messages
Failed to install Triton. 5 messages
Solved
Questions regarding submission for hiring challenge 6 messages
Open
Finetuning unsloth/Pixtral-12B-2409 vision dataset error 17 messages
How to know if QLoRA or LoRA is being applied 15 messages
Question about how performance is evaluated for puzzle A 2 messages
Multiple data type mismatches when finetuning Llama 3.2 Vision locally 2 messages
Open
Qwen-2.5-Coder-32B-Instruct fine-tuning training loss drops too fast 3 messages
Open
Training stops after first step 6 messages
Really quick question about problem B in unsloth challenges 2 messages
Qwen2.5-VL 7B Full Fine-tuning | CUDA out of memory | Colab Only 21 messages
Doing a QloRa cold start fine tuning before GRPO 3 messages
Clarification about Problem E in the Unsloth Puzzles 14 messages
Unsloth: Failed to create dynamic compiled modules! 2 messages
Open
RecursionError: maximum recursion depth exceeded 5 messages
Finetuning doesn't learn the LLM anything. 16 messages
Failed to create dynamic compiled after March updates 63 messages
Solved
Merging resets model performance LORA 2 messages
name error:name 'bias' is not defined 2 messages
NameError: name 'bias' is not defined 6 messages
Why the instruction is not masked with -100 in label during instruction fine tuning. 2 messages
Open
RuntimeError: Failed to find C compiler. Please specify via CC environment variable. 2 messages
Open
Kaggle for fine tuning Llama3.1_(8B)-GRPO.ipynb (I'm using the starter code) error when using peft 4 messages
Open
Unsloth example fintuning Colab failing when creating SFTTrainer 11 messages
CUDA out of memory for Custom Dataset [Qwen2-VL-7B-Instruct] 23 messages
Solved
finetune r1 distill model question 3 messages
Open
Attribute Error with SFTTrainer 7 messages
Out of Memory during GRPO training 9 messages
Practical Reinforcement Learning Examples Showcasing Unsloth's potential 262 messages
Open
fine tuning Llama-3.2-11B-Vision-Instruct with no success 5 messages
Qwen-2.5 GRPO finetune without vLLM 9 messages
vllm inference of lora finetuned unsloth model 10 messages
Code formatting of PRs for the `zoo` repository 9 messages
Issue with save_pretrained_merged 4 messages
Open
Load LORA Adapters and use for inference 9 messages
i want to fine tune the vision model.How can i prepare my owndataset 4 messages
OpenAI Compatible Server w/ DeepSeek-R1-dynamic-1.58-bit model 6 messages
Cannot Merge 4bit after GRPO 7 messages
Open