All issues

Search results

[serge] integration failure triage - 2026-08-07

Status: Open.

#47839 In huggingface/transformers;

· github-actions opened on Aug 7, 2026

additional_special_tokens is silently discarded when tokenizer_config.json also contains extra_special_tokens

bug

Status: Open.

#47838 In huggingface/transformers;

· wagnerpatriota opened on Aug 7, 2026

Qwen3_5GatedDeltaNet / Qwen3_5MoeGatedDeltaNet initialize A_log to -inf on some heads (bf16 models)

bug

Status: Open.

#47831 In huggingface/transformers;

· Nkluge-correa opened on Aug 7, 2026

Special tokens aren't escaped in template generation

bug

Status: Open.

#47822 In huggingface/transformers;

· pavelai opened on Aug 6, 2026

Whisper large-v3 and large-v3-turbo: config declares float16 but WhisperFeatureExtractor emits float32, so the manual path raises

Status: Open.

#47805 In huggingface/transformers;

· Kayvan-Zahiri opened on Aug 5, 2026

CVE-2025-14920 affected versions

Status: Open.

#47801 In huggingface/transformers;

· jesvinjames opened on Aug 5, 2026

Add RWKV-7 (Goose)

Status: Open.

#47787 In huggingface/transformers;

· Hakureirm opened on Aug 5, 2026

incorrect precedence of generation_config values

bug

Status: Open.

#47752 In huggingface/transformers;

· jkl-ds opened on Aug 3, 2026

HuBERT positional convolution is sensitive to padding when conv_pos_batch_norm=True

Status: Open.

#47739 In huggingface/transformers;

· Abdulrhman-shaheen opened on Aug 3, 2026

Add native support for OpenBMB MiniCPM4

New model

Status: Open.

#47732 In huggingface/transformers;

· sushmanthreddy opened on Aug 3, 2026

[Model Request] Native support for THUDM CogVLM2-Video

New model

Status: Open.

#47713 In huggingface/transformers;

· AryanSharma21 opened on Aug 2, 2026

RobertaTokenizerFast.from_pretrained() loads broken tokenizer in v5 (missing pre_tokenizer and BPE merges)

Status: Open.

#47706 In huggingface/transformers;

· alwaysprince05 opened on Aug 2, 2026