metadata
license: apache-2.0
datasets:
- Mielikki/Erebus-87k
- FourOhFour/Instruct_Phase
- FourOhFour/RP_Phase
- anthracite-core/full-opus-chosen-hermes-rejected-kto-v1
language:
- en
base_model:
- IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml
Aura-MoE-2x4B
Introduction
Aura-MoE-2x4B is a state of the art dedicated roleplaying model designed to fulfill your every desire.
The finetunes used in this merge saw several hundreds of millions of tokens of completion, instruction and roleplaying data. A Kahneman-Tversky Optimization was applied to both heal and give this model a unique output style.
This model can be considered inferior to Aura-MoE-2x4B-v2 which is a direct improvement.
Developed by Aura Industries, with contributions from Anthracite Org
Model Details
- Model Name: Aura-MoE-2x4B
- Base Model: IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml
- Model Type: Chat Completions
- Prompt Format: ChatML
- License: Apache-2.0
- Language: English
- Max Context: 8,192+ tokens
License
This model is licensed under the Apache 2.0 License.
Quantizations
Due to the abnormal nature of this model, only static GGUF quantization is available.
Open LLM Leaderboard Evaluation Results
Coming soon...
Metric | Value |
---|---|
Avg. | N/A |
IFEval (0-Shot) | N/A |
BBH (3-Shot) | N/A |
MATH Lvl 5 (4-Shot) | N/A |
GPQA (0-shot) | N/A |
MuSR (0-shot) | N/A |
MMLU-PRO (5-shot) | N/A |
Training Configuration
Click here for Mergekit and Axolotl configs
MoE Merge
base_model: FourOhFour/Crispy_Crab_4B
gate_mode: hidden
dtype: bfloat16
experts_per_token: 1
experts:
- source_model: FourOhFour/Crispy_Crab_4B
positive_prompts:
- "Roleplaying partner"
- source_model: FourOhFour/Zenith_4B
positive_prompts:
- "Instruction following assistant"
KTO
base_model: jeiku/2x4Bmoe
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
load_in_8bit: false
load_in_4bit: false
strict: false
hub_model_id: jeiku/moekto
hub_strategy: "all_checkpoints"
push_dataset_to_hub:
hf_use_auth_token: true
chat_template: chatml
rl: kto
rl_beta: 0.2
kto_desirable_weight: 0.2
datasets:
- path: anthracite-core/full-opus-chosen-hermes-rejected-kto-v1
type: chatml.argilla
shuffle_merged_datasets: true
val_set_size: 0.0
output_dir: ./outputs/out
sequence_len: 8192
sample_packing: false
eval_sample_packing: false
pad_to_sequence_len: false
wandb_project: moekto
wandb_entity:
wandb_watch:
wandb_name: moekto
wandb_log_model:
gradient_accumulation_steps: 16
micro_batch_size: 2
num_epochs: 2
max_steps: 500
optimizer: adamw_8bit
lr_scheduler: cosine
learning_rate: 0.00001
weight_decay: 0.05
train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: true
gradient_checkpointing: true
gradient_checkpointing_kwargs:
use_reentrant: true
remove_unused_columns: false
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
warmup_steps: 10
evals_per_epoch: 2
eval_table_size:
eval_max_new_tokens:
saves_per_epoch: 1
debug:
deepspeed:
fsdp:
fsdp_config:
fsdp:
fsdp_config:
special_tokens:
pad_token: <|finetune_right_pad_id|>