rbelanec/train_sst2_42_1779354537
The rbelanec/train_sst2_42_1779354537 model is a 1 billion parameter language model fine-tuned from meta-llama/Llama-3.2-1B-Instruct. It is specifically optimized for sentiment analysis tasks, having been fine-tuned on the SST-2 dataset. This model demonstrates a validation loss of 0.0908 on the SST-2 evaluation set, indicating its proficiency in binary sentiment classification.
Loading preview...
Overview
This model, rbelanec/train_sst2_42_1779354537, is a fine-tuned version of the meta-llama/Llama-3.2-1B-Instruct base model, specifically adapted for sentiment analysis. It has 1 billion parameters and was trained for a single epoch on the SST-2 dataset, achieving a validation loss of 0.0908.
Key Capabilities
- Sentiment Analysis: Specialized in classifying sentiment, as evidenced by its fine-tuning on the SST-2 dataset.
- Llama-3.2-1B-Instruct Base: Benefits from the foundational capabilities of the Llama-3.2-1B-Instruct architecture.
Training Details
The model was trained with a learning rate of 2e-06, a batch size of 8, and utilized the AdamW_Torch optimizer with a cosine learning rate scheduler. The training process involved 7201 steps, processing approximately 3.7 million input tokens.
Good For
- Binary Sentiment Classification: Ideal for tasks requiring the determination of positive or negative sentiment in text, particularly for short phrases or sentences similar to the SST-2 dataset.
- Research and Experimentation: Suitable for researchers exploring fine-tuning techniques on smaller Llama-based models for specific NLP tasks.