65 lines
1.2 KiB
YAML
65 lines
1.2 KiB
YAML
base_model: tiiuae/falcon-7b
|
|
trust_remote_code: true
|
|
model_type: AutoModelForCausalLM
|
|
tokenizer_type: AutoTokenizer
|
|
is_falcon_derived_model: true
|
|
load_in_8bit: true
|
|
load_in_4bit: false
|
|
gptq: false
|
|
strict: false
|
|
push_dataset_to_hub:
|
|
datasets:
|
|
- path: teknium/GPT4-LLM-Cleaned
|
|
type: alpaca:chat
|
|
dataset_prepared_path:
|
|
val_set_size: 0.05
|
|
adapter: lora
|
|
lora_model_dir:
|
|
sequence_len: 2048
|
|
max_packed_sequence_len:
|
|
lora_r: 16
|
|
lora_alpha: 32
|
|
lora_dropout: 0.0
|
|
lora_target_modules:
|
|
lora_target_linear: true
|
|
lora_fan_in_fan_out:
|
|
wandb_project:
|
|
wandb_entity:
|
|
wandb_watch:
|
|
wandb_run_id:
|
|
wandb_log_model:
|
|
output_dir: ./falcon-7b
|
|
batch_size: 2
|
|
micro_batch_size: 1
|
|
num_epochs: 4
|
|
optimizer: adamw_bnb_8bit
|
|
torchdistx_path:
|
|
lr_scheduler: cosine
|
|
learning_rate: 0.00003
|
|
train_on_inputs: false
|
|
group_by_length: false
|
|
bf16: true
|
|
fp16: false
|
|
tf32: true
|
|
gradient_checkpointing: true
|
|
early_stopping_patience:
|
|
resume_from_checkpoint:
|
|
local_rank:
|
|
logging_steps: 1
|
|
xformers_attention: true
|
|
flash_attention:
|
|
gptq_groupsize:
|
|
gptq_model_v1:
|
|
warmup_steps: 40
|
|
eval_steps: 5
|
|
save_steps: 43
|
|
debug:
|
|
deepspeed:
|
|
weight_decay: 0.0
|
|
fsdp:
|
|
fsdp_config:
|
|
special_tokens:
|
|
pad_token: "<|endoftext|>"
|
|
bos_token: ">>ABSTRACT<<"
|
|
eos_token: "<|endoftext|>"
|