Issue finetune
Hey, I'm trying to fine tune using Mixtral 8x22b on a A100.
The process keep getting stuck here : model-00010-of-00059.safetensors: 100%|_| 4.81G/4.81G [00:13<00:00, 344MB/s]
Downloading shards: 17%| | 10/59 [02:39<12:37, 15.46s/it^
Do you have any ideas why ?
Thanks :))))
base_model: mistralai/Mixtral-8x22B-Instruct-v0.1
model_type: MistralForCausalLM
tokenizer_type: AutoTokenizer
is_mistral_derived_model: true
hub_model_id: DevSelego/Mistral7b_summarizer_v9
load_in_8bit: false
load_in_4bit: true
strict: false
datasets:
- path: DevSelego/jobmaker_summarizer_v9
type: alpaca
dataset_prepared_path: last_run_prepared
val_set_size: 0.05
output_dir: ./qlora-out
adapter: qlora
lora_model_dir:
sequence_len: 8192
sample_packing: false
pad_to_sequence_len: true
lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:
lora_target_modules:
- gate_proj
- down_proj
- up_proj
- q_proj
- v_proj
- k_proj
- o_proj
wandb_project: Mistral_Jobmaker
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:
gradient_accumulation_steps: 1
micro_batch_size: 10
num_epochs: 1
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:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
loss_watchdog_threshold: 5.0
loss_watchdog_patience: 3
warmup_steps: 10
eval_steps: 0.05
eval_table_size:
eval_table_max_new_tokens: 128
save_steps:
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
bos_token: """
eos_token: "
unk_token: ""