smollm2-135m-instruct-paper-preference-150k-v1-sft-dpo-lr5e-6-ep1-beta0-1-lora16a32-seq1024
trained with verl for paper-query citation chunk grounding.
- base model:
HuggingFaceTB/SmolLM2-135M-Instruct - dataset:
paperbd/paper_preference_150K-v1 - training hyperparams:
sft-dpo-lr5e-6-ep1-beta0.1-lora16a32-seq1024 - local source folder:
paperhound
the dataset contains positive cited chunks, not the full arxiv paper haystack, so this model is trained to emit known supporting chunks for a paper/query pair.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for Pradheep1647/smollm2-135m-instruct-paper-preference-150k-v1-sft-dpo-lr5e-6-ep1-beta0-1-lora16a32-seq1024
Base model
HuggingFaceTB/SmolLM2-135M Quantized
HuggingFaceTB/SmolLM2-135M-Instruct