Issues

• Potential false negative issue in HardNegativeNLLLoss when hard negatives overlap with other samples' positives

  McGill-NLP/llm2vec#178 · Filter by author Gbone3176 opened on Apr 20, 2026

• Word-level tasks training

  McGill-NLP/llm2vec#175 · Filter by author Yayadongdong opened on Sep 24, 2025

• NaN in eval loss

  McGill-NLP/llm2vec#174 · Filter by author ultor1996 opened on Aug 27, 2025
  2

• How does config.json get created post MNTP training?

  McGill-NLP/llm2vec#173 · Filter by author karrtikiyer-tw opened on Aug 21, 2025

• support for GPT-oss 20b and 120b

  McGill-NLP/llm2vec#172 · Filter by author StephennFernandes opened on Aug 11, 2025

• Finetune recipe for classification on custom dataset

  McGill-NLP/llm2vec#171 · Filter by author vivian0429 opened on Jun 27, 2025

• Is it possible to train a decoder vec2llm?

  McGill-NLP/llm2vec#170 · Filter by author evgenytumanov opened on Jun 18, 2025
  1

• more dataset

  McGill-NLP/llm2vec#169 · Filter by author konioy opened on May 28, 2025

• split 1 and 2 in E5Data.py

  McGill-NLP/llm2vec#168 · Filter by author SiyuanHuangSJTU opened on Apr 17, 2025
  1

• CUDA Out of Memory Error When Loading LLM2Vec on GPU

  McGill-NLP/llm2vec#166 · Filter by author shrijayan opened on Mar 19, 2025

• Interesting idea

  McGill-NLP/llm2vec#165 · Filter by author kllgjc opened on Mar 10, 2025
  1

• Multi_GPU and AttributeError: Can't pickle local object 'add_hook_to_module.<locals>.new_forward

  McGill-NLP/llm2vec#164 · Filter by author pepper725 opened on Feb 26, 2025

• Huggingface: adaptor_model.safetensors

  McGill-NLP/llm2vec#163 · Filter by author pepper725 opened on Feb 23, 2025

• Can't find a checkpoint index (pytorch_model.bin.index.json or model.safetensors.index.json)

  McGill-NLP/llm2vec#162 · Filter by author Mem2019 opened on Feb 20, 2025

• Support with sentence transformers

  McGill-NLP/llm2vec#161 · Filter by author chrisconstant opened on Jan 10, 2025

• Activation layer

  McGill-NLP/llm2vec#160 · Filter by author arita37 opened on Dec 31, 2024

• Request for the toy data in Appendix F

  McGill-NLP/llm2vec#159 · Filter by author March-7 opened on Dec 23, 2024

• Guidance Needed: Adapting Pre-trained Llama Model for Image-to-Text Embedding with LLM2Vec and Unsloth

  McGill-NLP/llm2vec#157 · Filter by author LinMa910 opened on Dec 22, 2024

• How to Adapt the Supervised Contrastive Training for Custom Dataset?

  McGill-NLP/llm2vec#156 · Filter by author Franciscus-Carolus opened on Dec 16, 2024
  1

• Unable to Reproduce Results for Supervised Training on Echo Dataset with Mistral-7B-Instruct-v2

  McGill-NLP/llm2vec#154 · Filter by author ThonyPan opened on Nov 12, 2024
  8

• Introducing support for the IBM Granite 3.0 Model family.

  McGill-NLP/llm2vec#153 · Filter by author kailas711 opened on Oct 23, 2024

• NaN in Generated Embeddings

  McGill-NLP/llm2vec#152 · Filter by author superctj opened on Oct 20, 2024
  4

• Discrepancy in STS17 Results for S-LLaMA-1.3B

  McGill-NLP/llm2vec#151 · Filter by author newfish-lab opened on Oct 20, 2024
  3

• Issue with mntp training for Llama 3.2 model

  McGill-NLP/llm2vec#150 · Filter by author sandeep-krutrim opened on Oct 8, 2024

• Is it possible to resume from checkpoint in run_mntp.py?

  McGill-NLP/llm2vec#145 · Filter by author yallk opened on Sep 12, 2024
  1