Fix generate() to accept additional generation parameters

[Copilot]

Trainer.evaluate() passes length_penalty, no_repeat_ngram_size, and early_stopping to generate(), but SignLanguageTranslationModel.generate() didn't accept these parameters, causing a TypeError during training.

Changes

• Added **kwargs to SignLanguageTranslationModel.generate() signature
• Forward kwargs to underlying HuggingFace model for both encoder-decoder and causal LM paths

def generate(
    self,
    input_ids: torch.Tensor,
    attention_mask: torch.Tensor,
    max_length: int = 128,
    num_beams: int = 5,
    **kwargs,  # Now accepts length_penalty, early_stopping, etc.
):

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• huggingface.co
  • Triggering command: `/usr/bin/python python -c
    from transformers import AutoModelForSeq2SeqLM, AutoConfig, AutoTokenizer
    import inspect

Check generate method signature

model = AutoModelForSeq2SeqLM.from_pretrained('t5-small')
print('Generate parameters:')
sig = inspect.signature(model.generate)
fo` (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[Copilot]

Pull request overview

This pull request fixes a TypeError that occurred when Trainer.evaluate() called SignLanguageTranslationModel.generate() with additional generation parameters (length_penalty, no_repeat_ngram_size, and early_stopping) that the method didn't accept.

Changes:

• Added **kwargs parameter to the generate() method signature to accept additional generation parameters
• Forwarded kwargs to the underlying HuggingFace model's generate method for both encoder-decoder and causal LM code paths

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.