---
base_model: meta-llama/Llama-2-7b-hf
library_name: peft
---

# Reproduced RepLLaMA

This is a reproduced version of the RepLLaMA model.  See [this thread](https://github.com/texttron/tevatron/issues/129) for details of the reproduction process, which changed from their original version.

# Other Links
| Binary | Description |
|:-------|:------------|
| [samaya-ai/promptriever-llama2-7b-v1](https://huggingface.co/samaya-ai/promptriever-llama2-7b-v1) | A Promptriever bi-encoder model based on LLaMA 2 (7B parameters).|
| [samaya-ai/promptriever-llama3.1-8b-instruct-v1](https://huggingface.co/samaya-ai/promptriever-llama3.1-8b-instruct-v1) | A Promptriever bi-encoder model based on LLaMA 3.1 Instruct (8B parameters).|
| [samaya-ai/promptriever-llama3.1-8b-v1](https://huggingface.co/samaya-ai/promptriever-llama3.1-8b-v1) | A Promptriever bi-encoder model based on LLaMA 3.1 (8B parameters).|
| [samaya-ai/promptriever-mistral-v0.1-7b-v1](https://huggingface.co/samaya-ai/promptriever-mistral-v0.1-7b-v1) | A Promptriever bi-encoder model based on Mistral v0.1 (7B parameters). |
| [samaya-ai/RepLLaMA-reproduced](https://huggingface.co/samaya-ai/RepLLaMA-reproduced) | A reproduction of the RepLLaMA model (no instructions). A bi-encoder based on LLaMA 2, trained on the [tevatron/msmarco-passage-aug](https://huggingface.co/datasets/Tevatron/msmarco-passage-aug) dataset. |
| [samaya-ai/msmarco-w-instructions](https://huggingface.co/samaya-ai/msmarco-w-instructions) | A dataset of MS MARCO with added instructions and instruction-negatives, used for training the above models. |


# Usage

You can use this with the RepLLaMA example code in [tevatron](https://github.com/texttron/tevatron) or with mteb:

```python
import mteb
model = mteb.get_model("samaya-ai/RepLLaMA-reproduced")
tasks = mteb.get_tasks(tasks=["NFCorpus"], languages=["eng"])
evaluation = mteb.MTEB(tasks=tasks)
evaluation.run(model, batch_size=16)
```

The command used to create this reproduction was the Tevatron codebase (commit 9bb8381) with command:

```
#!/bin/bash
deepspeed --include localhost:0,1,2,3 --master_port 60000 --module tevatron.retriever.driver.train \
  --deepspeed deepspeed/ds_zero3_config.json \
  --output_dir retriever-llama2-4gpu \
  --model_name_or_path meta-llama/Llama-2-7b-hf \
  --lora \
  --lora_r 32 \
  --lora_target_modules q_proj,k_proj,v_proj,o_proj,down_proj,up_proj,gate_proj \
  --save_steps 200 \
  --dataset_name Tevatron/msmarco-passage-aug \
  --query_prefix "query: " \
  --passage_prefix "passage: " \
  --bf16 \
  --pooling eos \
  --append_eos_token \
  --normalize \
  --temperature 0.01 \
  --per_device_train_batch_size 8 \
  --gradient_checkpointing \
  --train_group_size 16 \
  --learning_rate 1e-4 \
  --query_max_len 32 \
  --passage_max_len 196 \
  --num_train_epochs 1 \
  --logging_steps 10 \
  --overwrite_output_dir \
  --warmup_steps 100 \
  --gradient_accumulation_steps 4
```


# Citation
For citation, please also see the [original RepLLaMA paper](https://arxiv.org/abs/2310.08319) and feel free to cite Promptriever as well:

```bibtex
@article{weller2024promptriever,
      title={Promptriever: Instruction-Trained Retrievers Can Be Prompted Like Language Models}, 
      author={Orion Weller and Benjamin Van Durme and Dawn Lawrie and Ashwin Paranjape and Yuhao Zhang and Jack Hessel},
      year={2024},
      eprint={2409.11136},
      archivePrefix={arXiv},
      primaryClass={cs.IR},
      url={https://arxiv.org/abs/2409.11136}, 
}
```