lyz

how to use lightseq inference engine

7

hello, I have 1 question about how to use lightseq inference engine > I trained an en2de model based on fairseq, which is a variant of the transformer, i.e. I...

[BUG]:

5

### 🐛 Describe the bug Error: failed to run torchrun --nproc_per_node=4 --nnodes=2 --node_rank=1 --rdzv_backend=c10d --rdzv_endpoint=ip:29501 --rdzv_id=colossalai-default-job train.py --strategy colossalai_zero2 on gpu-1648, is localhost: False, exception: No authentication methods available train...

[BUG]: how can i fine-tuning the glm-130b based on colossal-ai

1

### 🐛 Describe the bug hi, how can i fine-tuning the glm-130b model based on colossal-ai? glm-130b: https://keg.cs.tsinghua.edu.cn/glm-130b/zh/posts/glm-130b/ ### Environment _No response_

Hi, i am trying to convert Transformer model with tensorRT ops, the model is trained based on Opennmt-tf. here is my scripts: CUDA_VISIBLE_DEVICES=0 saved_model_cli convert --dir ./averaged-ende-export500k-v2 --tag_set serve --output_dir...

how to understand the code for calculating rewards

1

https://github.com/microsoft/DeepSpeedExamples/blob/master/applications/DeepSpeed-Chat/training/step3_rlhf_finetuning/ppo_trainer.py#L148 ``` def compute_rewards(self, prompts, log_probs, ref_log_probs, reward_score, action_mask): kl_divergence_estimate = -self.kl_ctl * (log_probs - ref_log_probs) rewards = kl_divergence_estimate start = prompts.shape[1] - 1 ends = start + action_mask[:, start:].sum(1)...