---
license: apache-2.0
library_name: peft
tags:
- axolotl
- generated_from_trainer
base_model: teknium/OpenHermes-2-Mistral-7B
model-index:
- name: Hermes-Agent
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
<details><summary>See axolotl config</summary>

axolotl version: `0.4.0`
```yaml
base_model: teknium/OpenHermes-2-Mistral-7B
base_model_config: teknium/OpenHermes-2-Mistral-7B
model_type: MistralForCausalLM
tokenizer_type: LlamaTokenizer
is_mistral_derived_model: true

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: THUDM/AgentInstruct
    type: sharegpt
    conversation: llama-2
    split: os
  - path: THUDM/AgentInstruct
    type: sharegpt
    conversation: llama-2
    split: db
  - path: THUDM/AgentInstruct
    type: sharegpt
    conversation: llama-2
    split: alfworld
  - path: THUDM/AgentInstruct
    type: sharegpt
    conversation: llama-2
    split: webshop
  - path: THUDM/AgentInstruct
    type: sharegpt
    conversation: llama-2
    split: kg
  - path: THUDM/AgentInstruct
    type: sharegpt
    conversation: llama-2
    split: mind2web

dataset_prepared_path: last_run_prepared
val_set_size: 0.1
output_dir: ./qlora-out

wandb_project: "Mistral-Agent"
wandb_log_model: "checkpoint"

hub_model_id: "gultar/Hermes-Agent"

chat_template: inst

adapter: qlora
lora_model_dir:

sequence_len: 8192
sample_packing: true
pad_to_sequence_len: true

lora_r: 8
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:
lora_target_modules:
  - gate_proj
  - down_proj
  - up_proj
  - q_proj
  - v_proj
  - k_proj
  - o_proj

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 1
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: false
group_by_length: false
bf16: true
fp16: false
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

loss_watchdog_threshold: 5.0
loss_watchdog_patience: 3

warmup_steps: 10
evals_per_epoch: 4
eval_table_size:
eval_table_max_new_tokens: 128
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
  bos_token: "<s>"
  eos_token: "</s>"
  unk_token: "<unk>"
lora_modules_to_save: 
  - lm_head
  - embed_tokens



```

</details><br>

# Hermes-Agent

This model is a fine-tuned version of [teknium/OpenHermes-2-Mistral-7B](https://huggingface.co/teknium/OpenHermes-2-Mistral-7B) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3002

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 10
- num_epochs: 1

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 0.6859        | 0.03  | 1    | 0.7320          |
| 0.3448        | 0.26  | 9    | 0.3382          |
| 0.4193        | 0.53  | 18   | 0.3233          |
| 0.2986        | 0.79  | 27   | 0.3002          |


### Framework versions

- PEFT 0.8.2
- Transformers 4.38.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.16.1
- Tokenizers 0.15.0