gemma-2b-tamil / README.md
abhinand's picture
Adding Evaluation Results (#1)
f8da377 verified
metadata
language:
  - en
  - ta
license: other
base_model: google/gemma-2b
datasets:
  - wikimedia/wikipedia
license_name: gemma-terms-of-use
license_link: https://ai.google.dev/gemma/terms
model-index:
  - name: gemma-2b-tamil
    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: 47.44
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abhinand/gemma-2b-tamil
          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: 71.3
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abhinand/gemma-2b-tamil
          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: 38.21
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abhinand/gemma-2b-tamil
          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: 34.93
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abhinand/gemma-2b-tamil
          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: 65.98
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abhinand/gemma-2b-tamil
          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: 12.89
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=abhinand/gemma-2b-tamil
          name: Open LLM Leaderboard

Gemma 2B Tamil v0.1 Alpha - Base Model [Experimental Release]

This is a Tamil foundational model continually pretrained from Google Gemma 2B. This is an experiment to see if Gemma can be adapted for Tamil without expanding vocabulary. While the responses may be rusty at times, it shows a lot of promise for a 2B parameter model.

Please Note: This model, labeled as a FOUNDATIONAL Language Model (LLM), is designed primarily for Causal Language Modeling (LM) purposes. In other words, if you are looking for an instruction following model in Tamil, you may find abhinand/gemma-2b-it-tamil-v0.1-alpha more suitable for your needs.

Procedure:

  1. The Gemma base model was continually pretrained on all available Tamil Wikipedia data for 3 epochs.
  2. The updated model was then finetuned on a mix of English and Tamil alpaca datasets for 5 epochs. Finetuned model can be found here.

Note: This project is currently under development (FOR TAMIL). The initial pretraining phase may not have been extensive enough, which suggests that the model's performance could improve by extending the pretraining on a larger dataset, such as CulturaX.

Model description

  • Model type: A 2B parameter GPT-like model continually pretrained on all available Tamil data from Wikipedia dataset.
  • Language(s): Bilingual. English and Tamil.
  • License: Google Gemma Terms of Use
  • Training Precision: bfloat16
  • Training Hardware: 4x Nvidia RTX 3090 GPUs
  • Training Cost: $20

Support my work

If you appreciate this work and would like to support its continued development, consider buying me a coffee. Your support is invaluable and greatly appreciated.

"Buy Me A Coffee"

Usage Note

It's important to note that the models have not undergone detoxification. Therefore, while they possess impressive linguistic capabilities, there is a possibility for them to generate content that could be deemed harmful or offensive. We urge users to exercise discretion and supervise the model's outputs closely, especially in public or sensitive applications.

Meet the Developers

Get to know the creators behind this innovative model and follow their contributions to the field:

We hope this model serves as a valuable tool in your NLP toolkit and look forward to seeing the advancements it will enable in the understanding and generation of the Tamil language.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric Value
Avg. 45.13
AI2 Reasoning Challenge (25-Shot) 47.44
HellaSwag (10-Shot) 71.30
MMLU (5-Shot) 38.21
TruthfulQA (0-shot) 34.93
Winogrande (5-shot) 65.98
GSM8k (5-shot) 12.89