Model Overview

laion/exp-uns-tezos-10x_glm_4_7_traces_jupiter_cleaned is an 8 billion parameter language model derived from the Qwen3-8B architecture. It has been specifically fine-tuned on a dataset identified as /data/cat/ws/befe330h-befe330h-otagent/huggingface/hub/datasets--DCAgent--exp-uns-tezos-10x_glm_4.7_traces_jupiter_cleaned/snapshots/2864d3bb974be2af999add0fb2c482c3605afc27_thinking_preprocessed.

Training Details

The model underwent training with the following key hyperparameters:

• Learning Rate: 4e-05
• Optimizer: ADAMW_TORCH_FUSED with betas=(0.9, 0.98) and epsilon=1e-08
• Batch Size: A total training batch size of 16 (train_batch_size: 1, gradient_accumulation_steps: 2 across 8 GPUs)
• Epochs: 7.0
• LR Scheduler: Cosine with a warmup ratio of 0.1

Intended Use Cases

While specific intended uses and limitations require further information, the fine-tuning on a dataset related to "tezos" and "traces" suggests its potential application in:

• Analyzing or generating content related to the Tezos blockchain.
• Processing or understanding trace data within the Tezos ecosystem.

Users should be aware that detailed information regarding its specific capabilities, performance benchmarks, and limitations is not yet available in the provided model description.