habanoz's picture
Adding Evaluation Results (#1)
4b74cd8 verified
metadata
language:
  - en
license: apache-2.0
datasets:
  - OpenAssistant/oasst_top1_2023-08-25
pipeline_tag: text-generation
model-index:
  - name: >-
      TinyLlama-1.1B-intermediate-step-715k-1.5T-lr-5-2.2epochs-oasst1-top1-instruct-V1
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: AI2 Reasoning Challenge (25-Shot)
          type: ai2_arc
          config: ARC-Challenge
          split: test
          args:
            num_few_shot: 25
        metrics:
          - type: acc_norm
            value: 31.48
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=habanoz/TinyLlama-1.1B-intermediate-step-715k-1.5T-lr-5-2.2epochs-oasst1-top1-instruct-V1
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: HellaSwag (10-Shot)
          type: hellaswag
          split: validation
          args:
            num_few_shot: 10
        metrics:
          - type: acc_norm
            value: 54.4
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=habanoz/TinyLlama-1.1B-intermediate-step-715k-1.5T-lr-5-2.2epochs-oasst1-top1-instruct-V1
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MMLU (5-Shot)
          type: cais/mmlu
          config: all
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 25.47
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=habanoz/TinyLlama-1.1B-intermediate-step-715k-1.5T-lr-5-2.2epochs-oasst1-top1-instruct-V1
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: TruthfulQA (0-shot)
          type: truthful_qa
          config: multiple_choice
          split: validation
          args:
            num_few_shot: 0
        metrics:
          - type: mc2
            value: 42.34
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=habanoz/TinyLlama-1.1B-intermediate-step-715k-1.5T-lr-5-2.2epochs-oasst1-top1-instruct-V1
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: Winogrande (5-shot)
          type: winogrande
          config: winogrande_xl
          split: validation
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 57.54
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=habanoz/TinyLlama-1.1B-intermediate-step-715k-1.5T-lr-5-2.2epochs-oasst1-top1-instruct-V1
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GSM8k (5-shot)
          type: gsm8k
          config: main
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 1.44
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=habanoz/TinyLlama-1.1B-intermediate-step-715k-1.5T-lr-5-2.2epochs-oasst1-top1-instruct-V1
          name: Open LLM Leaderboard

TinyLlama-1.1B-intermediate-step-715k-1.5T finetuned using OpenAssistant/oasst_top1_2023-08-25 dataset.

Qlora is used. Adapter is merged.

SFT code: https://github.com/habanoz/qlora.git

Command used:

accelerate launch $BASE_DIR/qlora/train.py \
  --model_name_or_path $BASE_MODEL \
  --working_dir $BASE_DIR/$OUTPUT_NAME-checkpoints \
  --output_dir $BASE_DIR/$OUTPUT_NAME-peft \
  --merged_output_dir $BASE_DIR/$OUTPUT_NAME \
  --final_output_dir $BASE_DIR/$OUTPUT_NAME-final \
  --num_train_epochs 2 \
  --logging_steps 1 \
  --save_strategy steps \
  --save_steps 75 \
  --save_total_limit 2 \
  --data_seed 11422 \
  --evaluation_strategy steps \
  --per_device_eval_batch_size 4 \
  --eval_dataset_size 0.01 \
  --eval_steps 75 \
  --max_new_tokens 1024 \
  --dataloader_num_workers 3 \
  --logging_strategy steps \
  --do_train \
  --do_eval \
  --lora_r 64 \
  --lora_alpha 16 \
  --lora_modules all \
  --bits 4 \
  --double_quant \
  --quant_type nf4 \
  --lr_scheduler_type constant \
  --dataset oasst1-top1 \
  --dataset_format oasst1 \
  --model_max_len 1024 \
  --per_device_train_batch_size 4 \
  --gradient_accumulation_steps 4 \
  --learning_rate 1e-5 \
  --adam_beta2 0.999 \
  --max_grad_norm 0.3 \
  --lora_dropout 0.0 \
  --weight_decay 0.0 \
  --seed 11422 \
  --gradient_checkpointing \
  --use_flash_attention_2 \
  --ddp_find_unused_parameters False

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric Value
Avg. 35.45
AI2 Reasoning Challenge (25-Shot) 31.48
HellaSwag (10-Shot) 54.40
MMLU (5-Shot) 25.47
TruthfulQA (0-shot) 42.34
Winogrande (5-shot) 57.54
GSM8k (5-shot) 1.44