Search results

Only tensorboard is installed without TensorFlow, causing undefined tf backend error

bug

Status: Open.

#45542 In huggingface/transformers;

· enze5088 opened on Apr 21, 2026

CLIPTokenizer uses 10**30 as model_max_length

bug

Status: Open.

#45538 In huggingface/transformers;

· D1-3105 opened on Apr 21, 2026

Feature Request: Add SCAO Optimizer integration for 1.5x faster fine-tuning throughput

Feature requestRequest for a new featureRequest for a new feature

Status: Open.

#45536 In huggingface/transformers;

· whispering3 opened on Apr 20, 2026

Add Olmo2ForSequenceClassification (and ideally OlmoForSequenceClassification / Olmo3ForSequenceClassification)

Status: Open.

#45529 In huggingface/transformers;

· earino opened on Apr 20, 2026

Feature request: Flash Attention 2 support for T5Gemma 2

Good Second IssueIssues that are more difficult to do than "Good First" issues - give it a try if you want!Issues that are more difficult to do than "Good First" issues - give it a try if you want!

Status: Open.

#45522 In huggingface/transformers;

· arunkumarchithanar opened on Apr 20, 2026

T5Gemma2: decoder self-attention fixed 4097-element mask at batch=1, fails on inputs >4094 tokens

Good Second IssueIssues that are more difficult to do than "Good First" issues - give it a try if you want!Issues that are more difficult to do than "Good First" issues - give it a try if you want!

Status: Open.

#45521 In huggingface/transformers;

· arunkumarchithanar opened on Apr 20, 2026

KeyError: 'flash_attn' in import_utils.py when running on Python 3.13

bug

Status: Open.

#45520 In huggingface/transformers;

· laichiaheng opened on Apr 20, 2026

Expose static_graph DDP flag via TrainingArguments

Status: Open.

#45518 In huggingface/transformers;

· KeitaW opened on Apr 20, 2026

GraniteMoEHybrid Model Calls Invalid Method

bug

Status: Open.

#45507 In huggingface/transformers;

· rnowling opened on Apr 18, 2026

Add V-JEPA 2.1 inference support

Feature requestRequest for a new featureRequest for a new feature

Status: Open.

#45496 In huggingface/transformers;

· davevanveen opened on Apr 17, 2026

[Gemma3] NaN embeddings on GPU when batching sequences of mixed length (sliding window attention + all-padding windows)

bug

Status: Open.

#45491 In huggingface/transformers;

· RiccardoTOTI opened on Apr 17, 2026

LlamaTokenizer in v5 overrides tokenizer.json's ByteLevel pre-tokenizer with Metaspace, silently breaks DeepSeek V3/R1 family

Status: Open.

#45488 In huggingface/transformers;

· dc3671 opened on Apr 17, 2026