nlpcat

support different pre_norm and layernorm_embedding in TransformerSpec for EncoderDecoderModel huggingface

2

can we make a change to support different prenorm and layernorm_embedding types for encoder and decoder in TransformerSpec for seq2seq model? Then it can support flexible encoder-decoder model in huggingface/transformers...

Would it be possible to add bert2gpt conversion? Thanks.

### Describe the feature allow huggingface.py mode to be passed in command cli. https://github.com/open-compass/opencompass/blob/9db56526382944560dde85b2b026dae82df6a2f7/opencompass/models/huggingface.py#L69 ### Will you implement it? - [ ] I would like to implement this feature and...

bug: alibi + multi_query_attention crash

2

```python try: import transformers except ImportError: pass from ctranslate2.specs import ( transformer_spec, ) from ctranslate2.converters.transformers import TransformersConverter, register_loader, ModelLoader, _SUPPORTED_ACTIVATIONS @register_loader("GPTBigCodeConfig") class GPTBigCodeMHALoader(ModelLoader): @property def architecture_name(self): return "GPTBigCodeForCausalLM" def get_model_spec(self,...

### System Info ```shell latest optimum version v1.14.1 ``` ### Who can help? _No response_ ### Information - [x] The official example scripts - [ ] My own modified scripts...