Zheng Yuan

Output inconsistent for autoregressive performer

2

I want to apply autoregressive performer for decoding. ```python import torch from performer_pytorch import Performer from torch import nn attn = Performer( dim = 128, heads = 4, depth =...

TODO

Hello, my work: CODER: Knowledge infused cross-lingual medical term embedding for term normalization (https://github.com/GanjinZero/CODER) (which is a contemporary work with your UmlSBERT) used confusing checkpoint name GanjinZero/UMLSBert_ENG in huggingface. Would...

We propose a new learning paradigm named RRHF (Rank Responses to Align Human Feedback) which does not need reinforcement learning and can perform on par with PPO to align human...

**Is your feature request related to a problem? Please describe.** We have posted a paper with codes [RRHF] (https://github.com/GanjinZero/RRHF) that can achieve human alignment without RLHF. RRHF needs 1-2 models...

Do you apply ablation study on wizardcoder with fine tuning with initial code alpaca?

2

In ablation study, you show multi round revision works better than single round revision. You also show wizardcoder outperforms Pretrained model starcoder. However, it is unknown how much wizardcoder benefits...

Also curious about more details on this work.

RRHF: Rank Responses to Align Language Models with Human Feedback without tears