Independent publisher
Liangzhi Li
Liangzhi-Li
Models in Library0
Datasets in Library2
Models on Hugging Face—
Followers—
Compare the published evidence
This profile connects the 2 model and dataset records shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Reported licenses: odc-by. Open a record to review its specific license and conditions.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.
Datasets
Training text, token files and checkpoints of small (124-million-parameter) Llama-architecture language models trained on FineWeb. Every checkpoint folder holds model.safetensors (float32) and config.json for transformers' LlamaForCausalLM: 12 layers, hidden size 768, 12 attention heads, MLP size 2048, context 2048, vocabulary 50,304 (the GPT-2 vocabulary padded), RMSNorm, rotary embeddings (θ = 10,000), input and output embeddings tied. Training of the base models: AdamW, learning rate 6e-4 with 300 warm-up steps and cosine decay to 6e-5, weight decay 0.1, gradient clipping 1.0, 524,288 tokens per step, 2,861 steps. FineWeb: ODC-By 1.0, subject to the Common Crawl terms of use. Cosmopedia…