Search results

[Bug] When loading LoRA via load_lora_adapter, the inference becomes very slow with high CPU usage

bug

Status: Open.

#3981 In unslothai/unsloth;

· elepherai opened on Feb 4, 2026

RuntimeError: Unsloth: Error patching SFTTrainer

Status: Open.

#3977 In unslothai/unsloth;

· Avvud opened on Feb 4, 2026

[Bug] Error while running GRPO notebook in ASUS gx10 machine (vLLM might be is the reason)

bug

Status: Open.

#3976 In unslothai/unsloth;

· damodharanj opened on Feb 3, 2026

finetuning Qwen3-30B-A3B-bnb-4bit with GRPO on A100

bug

Status: Open.

#3970 In unslothai/unsloth;

· amintavakoli1376 opened on Feb 3, 2026

GPT-oss-20b-BF16: torch.OutOfMemoryError: CUDA out of memory.

help wantedHelp from the OSS community wanted!Help from the OSS community wanted!

Status: Open.

#3966 In unslothai/unsloth;

· Decadz opened on Feb 3, 2026

CUDA error: CUBLAS_STATUS_EXECUTION_FAILED when calling cublasGemmEx (NVIDIA H100 NVL)

Status: Open.

#3963 In unslothai/unsloth;

· yenon118 opened on Feb 2, 2026

[Bug] RuntimeError: LRUCacheWorkerLoRAManager.create_lora_manager() takes 2 positional arguments but 3 were given

bug

Status: Open.

#3962 In unslothai/unsloth;

· haoyuhan1 opened on Feb 2, 2026

Request for Notebook to Fine-Tune Qwen TTS (or Alternatives Using Existing Notebooks)

feature requestFeature request pending on roadmapFeature request pending on roadmap

Status: Open.

#3961 In unslothai/unsloth;

· kailasas-supreme-ai opened on Feb 2, 2026

[Bug] RuntimeError: No config file found for Nemotron-3-Nano-30B-A3B on DGX Spark (NVIDIA GB10)

unsure bug?I'm unsureI'm unsure

Status: Open.

#3960 In unslothai/unsloth;

· SophieLin66 opened on Feb 2, 2026

How can I increase the GPU utilization while training / finetune an LLM with unsloth

help wantedHelp from the OSS community wanted!Help from the OSS community wanted!

Status: Open.

#3957 In unslothai/unsloth;

· dinusha94 opened on Feb 1, 2026

[Bug] SFTTrainer ignores bf16=False setting, causes error on T4 GPUs and Colab notebooks

bug

Status: Open.

#3956 In unslothai/unsloth;

· abrar360 opened on Jan 31, 2026

[DGX Spark / GB10] fast_inference=True disabled despite vLLM being present in Docker image

bug

Status: Open.

#3955 In unslothai/unsloth;

· hasso5703 opened on Jan 31, 2026