AmberYifan/capsdnum-marin-8b-base-math_random_b2000_s0

TEXT GENERATIONConcurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:8kTool Calling:SupportedPublished:Jul 16, 2026License:otherArchitecture:Transformer Featherless Exclusive Cold

The AmberYifan/capsdnum-marin-8b-base-math_random_b2000_s0 model is an 8 billion parameter language model, fine-tuned from marin-community/marin-8b-base. It was trained on the capsd_marin-8b-base-n80000-numina__mix_math_random_b2000_s0 dataset, suggesting a specialization in mathematical or numerical reasoning tasks. With an 8192-token context length, this model is suitable for applications requiring processing of moderately long sequences, particularly in domains related to its fine-tuning data.

Loading preview...

Model Overview

The AmberYifan/capsdnum-marin-8b-base-math_random_b2000_s0 is an 8 billion parameter language model, derived from the marin-community/marin-8b-base architecture. This model has been specifically fine-tuned using the capsd_marin-8b-base-n80000-numina__mix_math_random_b2000_s0 dataset, indicating a potential focus on numerical or mathematical reasoning tasks.

Training Details

The fine-tuning process involved specific hyperparameters:

  • Learning Rate: 1e-05
  • Batch Size: A train_batch_size of 2 and eval_batch_size of 8 were used, with a total_train_batch_size of 64 across 4 devices.
  • Optimizer: AdamW with betas=(0.9, 0.999) and epsilon=1e-08.
  • Scheduler: A cosine learning rate scheduler was employed with 0.03 warmup steps.
  • Epochs: The model was trained for 1 epoch.

Potential Use Cases

Given its fine-tuning on a dataset with "math_random" and "numina" in its name, this model is likely optimized for:

  • Mathematical problem-solving: Assisting with numerical calculations or mathematical reasoning.
  • Data analysis: Processing and interpreting numerical data.
  • Scientific computing: Generating or understanding content related to quantitative fields.

Finetunes
1 models