Search results

[Bug] --spec-type draft-mtp not present or being passed to LLAMA.cpp

bug

Status: Open.

#5535 In unslothai/unsloth;

· Carl-P4 opened on May 18, 2026

[Feature] MTP/Speculative Decoding - Disable model vision option

feature requestFeature request pending on roadmapFeature request pending on roadmap

Status: Open.

#5534 In unslothai/unsloth;

· Carl-P4 opened on May 18, 2026

[Bug] The following model_kwargs are not used by the model: ['backbone_last_hidden_state']

bug

Status: Open.

#5533 In unslothai/unsloth;

· fablevi opened on May 18, 2026

[Bug] Please fill in your issue title here.

bug

Status: Open.

#5532 In unslothai/unsloth;

· Jaskirat2504 opened on May 18, 2026

[Bug] MiMo-V2.5-GGUF UD-IQ3_S: llama.cpp loads after MiMo2 patches but outputs numeric garbage, likely GGUF/architecture mismatch

bug

Status: Open.

#5525 In unslothai/unsloth;

· cma-debug opened on May 17, 2026

[Feature] Change model parameters before loading the model

feature requestFeature request pending on roadmapFeature request pending on roadmap

Status: Open.

#5514 In unslothai/unsloth;

· melroy89 opened on May 17, 2026

[Feature] Decouple Tool Healing from Unsloth for use in external inference servers

feature requestFeature request pending on roadmapFeature request pending on roadmap

Status: Open.

#5502 In unslothai/unsloth;

· acerspyro opened on May 17, 2026

Feature Request: Publish Calibration Dataset for Dynamic-GGUF-Quants-2.0

feature requestFeature request pending on roadmapFeature request pending on roadmap

Status: Open.

#5500 In unslothai/unsloth;

· rikunarita opened on May 17, 2026

[Bug] save_pretrained_gguf() fails on Kaggle: ModuleNotFoundError: No module named 'conversion' (unsloth 2026.5.2)

bug

Status: Open.

#5495 In unslothai/unsloth;

· rajasingh012 opened on May 16, 2026

[Bug] Windows Studio llama-server runs CPU-only unless torch\lib is added to PATH for CUDA 13 DLLs

Status: Open.

#5491 In unslothai/unsloth;

· MilleniumGenAI opened on May 16, 2026

[Feature] Add max_concurrency setting for Unsloth API inference

feature requestFeature request pending on roadmapFeature request pending on roadmap

Status: Open.

#5481 In unslothai/unsloth;

· Yur3ka opened on May 16, 2026

[Bug] Gemma 4 handling multiple images in session incorrectly

bug

Status: Open.

#5470 In unslothai/unsloth;

· matix2267 opened on May 16, 2026