Training procedure
We finetuned mistralai/Mistral-7B-v0.1 on databricks/databricks-dolly-15k Dataset for 1 epoch using MonsterAPI no-code LLM finetuner.
Finetuning with MonsterAPI no-code LLM Finetuner in 5 easy steps:
- Select an LLM: Mistral 7B v0.1
- Select a task and Dataset: Instruction Finetuning and databricks-dolly-15k Dataset
- Specify Hyperparameters: We used default values suggested by finetuner
- Review and submit the job: That's it!
Hyperparameters & Run details:
- Model: mistralai/Mistral-7B-v0.1
- Dataset: databricks/databricks-dolly-15k
- Learning rate: 0.0002
- Number of epochs: 1
- Cutoff length: 512
- Data split: Training: 95% / Validation: 5%
- Gradient accumulation steps: 1
About Model:
The Mistral-7B-v0.1 Large Language Model (LLM) is a pretrained generative text model with 7 billion parameters. Mistral-7B-v0.1 outperforms Llama 2 13B on majority of the benchmarks as tested by Mistral team.
About Dataset:
databricks-dolly-15k is a corpus of more than 15,000 records generated by thousands of Databricks employees to enable large language models to exhibit the magical interactivity of ChatGPT.
Framework versions
- PEFT 0.5.0
