delayedkarma/mistral-7b-text-to-sql_full-model

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

mistral-7b-text-to-sql_full-model

• —This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the b-mc2/sql-create-context dataset.
• —These are the full model weights (merged with adapter weights), and the code to use these for generation is given below.
• —Primary reference: https://www.philschmid.de/fine-tune-llms-in-2024-with-trl

Model description

• —Model type: Language model
• —Language(s) (NLP): English
• —License: Apache 2.0
• —Finetuned from model : Mistral-7B-v0.1

How to get started with the model

import torch

from datasets import load_dataset
from transformers import AutoTokenizer, AutoModelForCausalLM

# Load model directly

tokenizer = AutoTokenizer.from_pretrained("delayedkarma/mistral-7b-text-to-sql_full-model")
model = AutoModelForCausalLM.from_pretrained("delayedkarma/mistral-7b-text-to-sql_full-model")

text = "How many matched scored 3–6, 7–6(5), 6–3?"
inputs = tokenizer(text, return_tensors="pt")

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• —learning_rate: 0.0002
• —trainbatchsize: 3
• —evalbatchsize: 8
• —seed: 42
• —gradientaccumulationsteps: 2
• —totaltrainbatch_size: 6
• —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• —lrschedulertype: constant
• —lrschedulerwarmup_ratio: 0.03
• —num_epochs: 3

Framework versions

• —PEFT 0.7.2.dev0
• —Transformers 4.36.2
• —Pytorch 2.2.2
• —Datasets 2.16.1
• —Tokenizers 0.15.2