Issues

• [Bug] In unified training with negatives_cross_device, the embedding/generation loss balance depends on the number of GPUs

  ContextualAI/gritlm#75 · Filter by author gss10282025 opened on Sep 26, 2026
  1

• Severe performance degradation during Mistral fine-tuning

  ContextualAI/gritlm#73 · Filter by author sdoubleoj opened on Jun 14, 2025
  5

• Potential Inconsistencies Between Repo and Model License

  ContextualAI/gritlm#72 · Filter by author yueyangchen1 opened on May 14, 2025

• How to train using hard negative only?

  ContextualAI/gritlm#71 · Filter by author zcakzhuu opened on Jan 22, 2025
  4

• saving the trained model for inference

  ContextualAI/gritlm#66 · Filter by author zhj2022 opened on Nov 20, 2024
  1

• CUDA OOM when finetuning meta-llama/Meta-Llama-3-8B-Instruct

  ContextualAI/gritlm#65 · Filter by author zhj2022 opened on Nov 20, 2024
  1

• Train llama 3.1 with GRIT

  ContextualAI/gritlm#60 · Filter by author ThisisXXZ opened on Nov 9, 2024
  5

• Evaluate GritLM-7B on MTEB datasets

  ContextualAI/gritlm#57 · Filter by author ThisisXXZ opened on Nov 4, 2024
  9

• how to add projection

  ContextualAI/gritlm#53 · Filter by author threestone965 opened on Sep 10, 2024
  1

• Can this training framework be used for training gte-qwen2-7b-instruct model？

  ContextualAI/gritlm#51 · Filter by author Double-bear opened on Aug 28, 2024
  1

• save checkpoint with error

  ContextualAI/gritlm#50 · Filter by author 13613979212 opened on Aug 28, 2024
  1

• question about emb loss

  ContextualAI/gritlm#49 · Filter by author 256785 opened on Jul 24, 2024
  3

• TypeError: TextEncodeInput must be Union[TextInputSequence, Tuple[InputSequence, InputSequence]]

  ContextualAI/gritlm#48 · Filter by author bo-jpg opened on Jul 24, 2024
  6

• sentence transforemr

  ContextualAI/gritlm#47 · Filter by author drewskidang opened on Jul 5, 2024
  2

• About OOM During Training and Questions Regarding Attn

  ContextualAI/gritlm#46 · Filter by author Zheng-Jay opened on Jul 2, 2024
  7

• Can I select causal attention for retrieval embeddings when using GritLM

  ContextualAI/gritlm#45 · Filter by author Yangseung opened on Jul 2, 2024
  17

• When training a unified model, TypeError: MistralForCausalLM.forward() got an unexpected keyword argument 'is_causal'

  ContextualAI/gritlm#41 · Filter by author zillion-zhao opened on Jun 14, 2024
  5

• How do I get the sentence embedding from GritLM/emb_m7_nodes16_fast?

  ContextualAI/gritlm#36 · Filter by author phartman-keysight opened on May 24, 2024
  4

• projection layer

  ContextualAI/gritlm#35 · Filter by author CementMaker opened on May 24, 2024
  2

• CustomRandomSampler not working in huggingface Trainer and Accelerator

  ContextualAI/gritlm#31 · Filter by author YanshekWoo opened on Apr 25, 2024
  1

• Unable to load trained model

  ContextualAI/gritlm#27 · Filter by author deepakkr-singh opened on Apr 13, 2024
  1

• Is there hope Grit-Embedding beats this task?

  ContextualAI/gritlm#22 · Filter by author marioeljuga opened on Apr 3, 2024
  2

• E5 dataset

  ContextualAI/gritlm#17 · Filter by author wangskyGit opened on Mar 20, 2024
  1

• RuntimeError

  ContextualAI/gritlm#16 · Filter by author BlackHandsomeLee opened on Mar 17, 2024
  1

• bidirectional attention or casual attention for embedding?

  ContextualAI/gritlm#15 · Filter by author yonxie opened on Mar 15, 2024
  5