# Qwen/Qwen1.5-7B

base_model: Qwen/Qwen1.5-7B
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer

# is_qwen_derived_model: true
trust_remote_code: true

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: OdiaGenAIdata/culturax-odia
    type: completion
dataset_prepared_path:
val_set_size: 0.05
output_dir: ./lora-out-qwen-7b-odia
hub_model_id: sam2ai/qwen_1.5_odia_7b

sequence_len: 2048  # supports up to 8192
sample_packing: false
pad_to_sequence_len:

adapter: qlora
lora_model_dir:
lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:

wandb_project: Qwen-completion-7b-odia
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 2
micro_batch_size: 1
num_epochs: 10
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: false
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention:

warmup_steps: 10
evals_per_epoch: 4
eval_table_size:
eval_table_max_new_tokens: 128
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:

qwen_1.5_odia_7b

This model is a fine-tuned version of Qwen/Qwen1.5-7B on the None dataset. It achieves the following results on the evaluation set:

Loss: nan

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.0002
train_batch_size: 1
eval_batch_size: 1
seed: 42
distributed_type: multi-GPU
num_devices: 8
gradient_accumulation_steps: 2
total_train_batch_size: 16
total_eval_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 10
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss
0.8734	0.0	1	nan
0.377	0.25	5463	nan
0.3727	0.5	10926	nan
0.3894	0.75	16389	nan
0.3933	1.0	21852	nan
0.3542	1.25	27315	nan
0.3567	1.5	32778	nan
0.3557	1.75	38241	nan
0.3414	2.0	43704	nan
0.3261	2.25	49167	nan
0.3813	2.5	54630	nan
0.3607	2.75	60093	nan
0.3348	3.0	65556	nan
0.3464	3.25	71019	nan
0.3545	3.5	76482	nan
0.2719	3.75	81945	nan
0.3158	4.0	87408	nan
0.3119	4.25	92871	nan
0.3311	4.5	98334	nan
0.3335	4.75	103797	nan
0.3399	5.0	109260	nan
0.3247	5.25	114723	nan
0.3166	5.5	120186	nan
0.3366	5.75	125649	nan
0.3478	6.0	131112	nan
0.2852	6.25	136575	nan
0.2852	6.5	142038	nan
0.2601	6.75	147501	nan
0.2734	7.0	152964	nan
0.2983	7.25	158427	nan
0.2068	7.5	163890	nan
0.2355	7.75	169353	nan
0.2836	8.0	174816	nan
0.2263	8.25	180279	nan
0.2953	8.5	185742	nan
0.257	8.75	191205	nan
0.2484	9.0	196668	nan
0.2477	9.25	202131	nan
0.2641	9.5	207594	nan
0.2569	9.75	213057	nan
0.2846	10.0	218520	nan

Framework versions

PEFT 0.8.2
Transformers 4.37.0
Pytorch 2.0.1+gita61a294
Datasets 2.16.1
Tokenizers 0.15.0