Closed manishiitg closed 8 months ago
this error doesn't come when i used A100X4 GPU's but only comes when is used A100 x 8 GPU's
Could you try upgrading deepspeed according to https://github.com/OpenAccess-AI-Collective/axolotl/blob/main/docs/faq.md
Try the docker image with these args
sudo docker run --gpus '"all"' --ipc=host --ulimit memlock=-1 --ulimit stack=67108864 --rm -it winglian/axolotl:main-py3.10-cu118-2.0.1
above worked :)
Please check that this issue hasn't been reported before.
Expected Behavior
should work
Current behaviour
Steps to reproduce
i have tried using deepspeed zero1, zero3, zero2 all
Config yaml
base_model: mistralai/Mistral-7B-v0.1 model_type: MistralForCausalLM tokenizer_type: LlamaTokenizer is_mistral_derived_model: true
load_in_8bit: false load_in_4bit: true strict: false
datasets:
hub_model_id: manishiitg/Mistral-databricks-databricks-dolly-15k-hi
hf_use_auth_token: true
dataset_prepared_path: val_set_size: 0.05 output_dir: /sky-notebook/manishiitg/Mistral-databricks-databricks-dolly-15k-hi
adapter: qlora lora_model_dir:
sequence_len: 4096 sample_packing: true pad_to_sequence_len: true
lora_r: 16 lora_alpha: 32 lora_dropout: 0.05 lora_target_linear: true lora_fan_in_fan_out: lora_target_modules:
wandb_project: wandb_entity: wandb_watch: wandb_run_id: wandb_log_model:
gradient_accumulation_steps: 4 micro_batch_size: 8 num_epochs: 2 optimizer: adamw_bnb_8bit lr_scheduler: cosine learning_rate: 0.0002
train_on_inputs: false group_by_length: false bf16: true fp16: false tf32: false
gradient_checkpointing: true early_stopping_patience: resume_from_checkpoint: auto_resume_from_checkpoints: true ## manage check point resume from here local_rank: logging_steps: 1 xformers_attention: flash_attention: true
warmup_steps: 10 eval_steps: 0.05 eval_table_size: eval_table_max_new_tokens: 128 save_steps: 100 ## increase based on your dataset save_strategy: steps debug: deepspeed: /sky_workdir/zero3.json weight_decay: 0.0 fsdp: fsdp_config: special_tokens: bos_token: ""
" eos_token: "" unk_token: "Possible solution
No response
Which Operating Systems are you using?
Python Version
3.10
axolotl branch-commit
main
Acknowledgements