nous-gemma-seven

<!-- 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"/>

<details><summary>See axolotl config</summary>

axolotl version: 0.4.0

base_model: MesozoicMetallurgist/nous-Induan
model_type: GemmaForCausalLM
hub_model_id: nous-gemma-seven

load_in_8bit: false
load_in_4bit: false
strict: false


datasets:
  - path: tomaszki/gemma
  - path: tomaszki/gemma-1
  - path: tomaszki/gemma-2
  - path: tomaszki/gemma-3
  - path: tomaszki/gemma-4
  - path: tomaszki/gemma-5
  - path: tomaszki/gemma-6
  - path: tomaszki/gemma-7
  - path: tomaszki/gemma-8
  - path: tomaszki/gemma-9
  - path: tomaszki/gemma-10
  - path: tomaszki/gemma-11
val_set_size: 0.0
output_dir: out

sequence_len: 1024
sample_packing: false

wandb_project: axolotl
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 30
micro_batch_size: 7
num_epochs: 1
optimizer: adamw_hf
lr_scheduler: cosine
learning_rate: 0.0000005
cosine_min_lr_ratio: 0.5
max_grad_norm: 0.00001

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

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

warmup_steps: 0
saves_per_epoch: 1
debug:
deepspeed: #deepspeed_configs/zero2.json
weight_decay: 0.01
fsdp:
fsdp_config:
special_tokens:

</details><br>

nous-gemma-seven

This model is a fine-tuned version of MesozoicMetallurgist/nous-Induan on the None dataset.

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: 5e-07
• train_batch_size: 7
• eval_batch_size: 7
• seed: 42
• gradient_accumulation_steps: 30
• total_train_batch_size: 210
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 1

Training results

Framework versions

• Transformers 4.39.0.dev0
• Pytorch 2.1.2+cu118
• Datasets 2.17.1
• Tokenizers 0.15.0