miiikiik/Qwen3-8B-music-movie-coa-sft

Model Overview

The miiikiik/Qwen3-8B-music-movie-coa-sft is an 8 billion parameter language model, fine-tuned from an existing Qwen3-8B variant that was previously adapted for music-related tasks. This iteration has undergone additional supervised fine-tuning specifically on the movie_coa_sft dataset, indicating a specialization in movie-related content and potentially a combination of music and movie contexts.

Key Characteristics

• Base Model: Qwen3-8B architecture.
• Parameter Count: 8 billion parameters.
• Context Length: Supports a substantial context window of 32768 tokens.
• Specialization: Fine-tuned on the movie_coa_sft dataset, building upon prior music-related fine-tuning.

Training Details

The model was trained using the following hyperparameters:

• Learning Rate: 1e-05
• Batch Size: train_batch_size of 2, eval_batch_size of 8, with a gradient_accumulation_steps of 4, resulting in a total_train_batch_size of 64.
• Optimizer: ADAMW_TORCH with default betas and epsilon.
• Scheduler: Cosine learning rate scheduler with a 0.1 warmup ratio.
• Epochs: 3.0 epochs.

Intended Use Cases

This model is particularly suited for applications requiring detailed understanding, generation, or analysis within the music and movie domains, especially those benefiting from its specific fine-tuning on movie-related conversational or analytical data. Its large context window allows for processing longer inputs and generating more coherent and contextually relevant outputs in these areas.