OptimalScale/LMFlow

Issues

• [DPO is available?]

  #741 opened 9 months ago
  2

• Running install.sh after git clone requires over 200GB Ram

  #732 opened 6 months ago
  6

• Multiple rounds of training

  #731 opened 7 months ago
  1

• is support llava model ?

  #730 opened 9 months ago
  3

• Question Regarding Optimizer States Management in Lisa Implementation

  #729 opened 9 months ago
  5

• [BUG] LISA: same loss regardless of lisa_activated_layers

  #726 opened 9 months ago
  17

• [BUG] LISA Finetune: AttributeError: 'ChatGLMModel' object has no attribute 'h'

  #719 opened 9 months ago
  3

• Question

  #714 opened 9 months ago
  8

• [BUG] Conflict dependencies of package versions

  #695 opened 9 months ago
  1

• How to verify the training is successful

  #694 opened a year ago
  3

• [BUG] deepspeed.runtime.zero.utils.ZeRORuntimeException: You are using ZeRO-Offload with a client provided optimizer

  #693 opened a year ago
  2

• [BUG]Output my input

  #692 opened a year ago
  4

• Perform Inference using Python instead of command line

  #691 opened a year ago
  1

• finetune ChatGLM with lora[BUG]

  #690 opened a year ago
  2

• Use Ollama to perform inference with trained model using LMFLOW

  #689 opened a year ago
  1

• AttributeError: 'DeepSpeedCPUAdam' object has no attribute 'ds_opt_adam'

  #688 opened a year ago
  1

• Minimum Requirement of GPU for Fine Tuning

  #687 opened a year ago
  4

• [BUG] main branch deepspeed error report

  #686 opened a year ago
  1

• the problem of scripts

  #685 opened a year ago
  1

• Distributed training parameter settings

  #684 opened a year ago
  1

• No checkpoints saved with full finetune even when I set --save_steps

  #683 opened a year ago
  1

• Training 70b model

  #682 opened a year ago
  1

• Experiments for speculative_decoding

  #680 opened a year ago
  3

• baichuan-inc_Baichuan2-7B-Chat can't training

  #678 opened a year ago
  2

• LoRA + FlashAttention2 speed up？

  #677 opened a year ago
  1

• How to fix the git clone

  #676 opened a year ago
  1

• [BUG] constant loss during fine-tuning

  #675 opened a year ago
  4

• [BUG]./scripts/run_finetune_with_lora_save_aggregated_weights.sh

  #672 opened a year ago
  3

• [Request For New Feature] Support InternLM Model

  #671 opened a year ago
  1

• [BUG]Loading a dataset cached in a LocalFileSystem is not supported

  #666 opened a year ago
  2

• A6000 support for FlashAttentionV2

  #665 opened a year ago
  3

• [BUG] Locally finetune failed

  #664 opened a year ago
  2

• [BUG] After finetune, model generate useless tokens.[全量微调完，模型输出连续<sep>]

  #663 opened a year ago
  1

• [BUG]Model size change

  #662 opened a year ago
  5

• lora training - Input：Instruction

  #661 opened a year ago
  1

• Issues of local deployment

  #660 opened a year ago
  3

• Issue on default temperature settings

  #659 opened a year ago
  1

• baichuan7B-2 trust_remote_code and then followed by other Error

  #655 opened a year ago
  9

• Support Mistral 7B model

  #652 opened a year ago
  3

• what's the difference between text_only and text2text data type?

  #645 opened a year ago
  8

• [BUG]'RaftTrainer' object has no attribute 'accelerator'

  #644 opened a year ago
  1

• [BUG]deepspeed_init() got an unexpected keyword argument 'resume_from_checkpoint'

  #643 opened a year ago
  0

• The raft problem,when I tried to run "./scripts/run_raft_align.sh"

  #639 opened a year ago
  4

• module named 'lmflow.args'

  #637 opened a year ago
  7

• I have solved this issue,please remove this issue.thanks.

  #632 opened a year ago
  0

• can't training codellama

  #631 opened a year ago
  2

• Fine-Tuning llama2-7b-chat model

  #629 opened a year ago
  1

• The Vocab size mismatch[BUG]

  #627 opened a year ago
  0

• About the eos token

  #626 opened a year ago
  4

• abnormal checkpoint

  #623 opened a year ago
  1