metadata

license: other
base_model: Qwen/Qwen1.5-4B
tags:
  - generated_from_trainer
datasets:
  - tyzhu/lmind_hotpot_train8000_eval7405_v1_reciteonly_qa
metrics:
  - accuracy
model-index:
  - name: >-
      lmind_hotpot_train8000_eval7405_v1_reciteonly_qa_Qwen_Qwen1.5-4B_5e-4_lora2
    results:
      - task:
          name: Causal Language Modeling
          type: text-generation
        dataset:
          name: tyzhu/lmind_hotpot_train8000_eval7405_v1_reciteonly_qa
          type: tyzhu/lmind_hotpot_train8000_eval7405_v1_reciteonly_qa
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.6595924308588064
library_name: peft

lmind_hotpot_train8000_eval7405_v1_reciteonly_qa_Qwen_Qwen1.5-4B_5e-4_lora2

This model is a fine-tuned version of Qwen/Qwen1.5-4B on the tyzhu/lmind_hotpot_train8000_eval7405_v1_reciteonly_qa dataset. It achieves the following results on the evaluation set:

Loss: 2.6245
Accuracy: 0.6596

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: 0.0005
train_batch_size: 1
eval_batch_size: 2
seed: 42
distributed_type: multi-GPU
num_devices: 4
gradient_accumulation_steps: 8
total_train_batch_size: 32
total_eval_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: constant
lr_scheduler_warmup_ratio: 0.05
num_epochs: 20.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.4465	1.0	250	1.4944	0.6773
1.2673	2.0	500	1.5207	0.6758
1.0289	3.0	750	1.6096	0.6728
0.8374	4.0	1000	1.7156	0.6700
0.6351	5.0	1250	1.8476	0.6668
0.5167	6.0	1500	1.9508	0.6650
0.3838	7.0	1750	2.0580	0.6636
0.3269	8.0	2000	2.1726	0.6622
0.2495	9.0	2250	2.2331	0.6618
0.2321	10.0	2500	2.3444	0.6616
0.1914	11.0	2750	2.3785	0.6613
0.1895	12.0	3000	2.4416	0.6613
0.1628	13.0	3250	2.4870	0.6601
0.1679	14.0	3500	2.4965	0.6600
0.1478	15.0	3750	2.5362	0.6606
0.1571	16.0	4000	2.5291	0.6608
0.1385	17.0	4250	2.5737	0.6600
0.146	18.0	4500	2.5954	0.6594
0.1335	19.0	4750	2.6082	0.6596
0.1438	20.0	5000	2.6245	0.6596

Framework versions

PEFT 0.5.0
Transformers 4.40.2
Pytorch 2.3.0
Datasets 2.19.1
Tokenizers 0.19.1