---
library_name: transformers
base_model: Dans-DiscountModels/Mistral-Nemo-Base-2407-ChatML-Mod
tags:
- axolotl
- generated_from_trainer
datasets:
- PocketDoc/Dans-Reasoningmaxx-NaturalReasoning
- PocketDoc/Dans-Reasoningmaxx-WebInstruct
- PocketDoc/Dans-Benchmaxx-COT
- PocketDoc/Dans-Logicmaxx-SAT-AP
- PocketDoc/Dans-Assistantmaxx-Opus-Merge
- PocketDoc/Dans-Assistantmaxx-sonnetorca-subset
model-index:
- name: 12b-mn-dans-reasoning-test-4
results: []
---
[
](https://github.com/axolotl-ai-cloud/axolotl)
See axolotl config
axolotl version: `0.8.0.dev0`
```yaml
base_model: Dans-DiscountModels/Mistral-Nemo-Base-2407-ChatML-Mod
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
trust_remote_code:
# wandb configuration
wandb_project: 12b-mn-dans-reasoning-test
wandb_watch:
wandb_run_id: V0.0.3-1-3 # V{Version}-{Run Number}-{Attempt Number}
wandb_log_model:
# push checkpoints to hub
hub_model_id: Dans-DiscountModels/12b-mn-dans-reasoning-test-4
# how to push checkpoints to hub
# https://huggingface.co/docs/transformers/v4.31.0/en/main_classes/trainer#transformers.TrainingArguments.hub_strategy
hub_strategy: "every_save"
# Whether to use hf `use_auth_token` for loading datasets. Useful for fetching private datasets
# Required to be true when used in combination with `push_dataset_to_hub`
hf_use_auth_token: true
# where to save the finished model to
output_dir: ./12b-mn-dans-reasoning-test
save_safetensors: true
# dataset settings (local or huggingface repo)
datasets:
- path: PocketDoc/Dans-Reasoningmaxx-NaturalReasoning
type: dan-chat-advanced
- path: PocketDoc/Dans-Reasoningmaxx-WebInstruct
type: dan-chat-advanced
- path: PocketDoc/Dans-Benchmaxx-COT
type: dan-chat-advanced
- path: PocketDoc/Dans-Logicmaxx-SAT-AP
type: dan-chat-advanced
- path: PocketDoc/Dans-Assistantmaxx-Opus-Merge
type: dan-chat-advanced
- path: PocketDoc/Dans-Assistantmaxx-sonnetorca-subset
type: dan-chat-advanced
plugins:
- axolotl.integrations.liger.LigerPlugin
- axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin
liger_rope: true
liger_rms_norm: true
liger_layer_norm: true
liger_glu_activation: true
liger_fused_linear_cross_entropy: false
cut_cross_entropy: true
load_in_8bit: false
load_in_4bit: false
strict: false
adapter:
lora_model_dir:
lora_r: 128
lora_alpha: 128
lora_dropout: 0.1
lora_target_linear: True
lora_target_modules:
lora_modules_to_save:
- embed_tokens
- lm_head
lora_fan_in_fan_out:
peft_use_rslora: true
dataset_prepared_path: ./12b-mn-dans-reasoning-test-data
val_set_size: 0.005
sequence_len: 8192
sample_packing: true
eval_sample_packing: true
pad_to_sequence_len: true
gradient_checkpointing: true
# gradient_checkpointing_kwargs:
# use_reentrant: false
gradient_accumulation_steps: 1
micro_batch_size: 4
num_epochs: 2
optimizer: came_pytorch
lr_scheduler: rex
learning_rate: 0.0000015
cosine_min_lr_ratio: 0.1
weight_decay: 0.1
max_grad_norm: 0.1
train_on_inputs: false
group_by_length: true
bf16: true
fp16: false
tf32: false
early_stopping_patience:
resume_from_checkpoint:
auto_resume_from_checkpoints: true
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
warmup_ratio: 0.05
evals_per_epoch: 16
eval_table_size:
eval_max_new_tokens:
saves_per_epoch: 8
save_total_limit: 1
debug: false
deepspeed: deepspeed_configs/zero3_bf16.json
fsdp:
fsdp_config:
special_tokens:
```
# 12b-mn-dans-reasoning-test-4
This model is a fine-tuned version of [Dans-DiscountModels/Mistral-Nemo-Base-2407-ChatML-Mod](https://huggingface.co/Dans-DiscountModels/Mistral-Nemo-Base-2407-ChatML-Mod) on the PocketDoc/Dans-Reasoningmaxx-NaturalReasoning, the PocketDoc/Dans-Reasoningmaxx-WebInstruct, the PocketDoc/Dans-Benchmaxx-COT, the PocketDoc/Dans-Logicmaxx-SAT-AP, the PocketDoc/Dans-Assistantmaxx-Opus-Merge and the PocketDoc/Dans-Assistantmaxx-sonnetorca-subset datasets.
It achieves the following results on the evaluation set:
- Loss: 0.6204
## 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: 1.5e-06
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- total_train_batch_size: 16
- total_eval_batch_size: 16
- optimizer: Use OptimizerNames.ADAMW_HF with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 38
- num_epochs: 2.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 0.8049 | 0.0026 | 1 | 0.8338 |
| 0.8006 | 0.0644 | 25 | 0.7551 |
| 0.7197 | 0.1289 | 50 | 0.7009 |
| 0.7165 | 0.1933 | 75 | 0.6810 |
| 0.7183 | 0.2577 | 100 | 0.6697 |
| 0.6671 | 0.3222 | 125 | 0.6620 |
| 0.6406 | 0.3866 | 150 | 0.6567 |
| 0.6656 | 0.4510 | 175 | 0.6528 |
| 0.6539 | 0.5155 | 200 | 0.6489 |
| 0.634 | 0.5799 | 225 | 0.6460 |
| 0.6606 | 0.6443 | 250 | 0.6428 |
| 0.6815 | 0.7088 | 275 | 0.6401 |
| 0.6082 | 0.7732 | 300 | 0.6385 |
| 0.6754 | 0.8376 | 325 | 0.6364 |
| 0.6284 | 0.9021 | 350 | 0.6347 |
| 0.6517 | 0.9665 | 375 | 0.6326 |
| 0.5583 | 1.0309 | 400 | 0.6340 |
| 0.5716 | 1.0954 | 425 | 0.6328 |
| 0.5799 | 1.1598 | 450 | 0.6323 |
| 0.5957 | 1.2242 | 475 | 0.6316 |
| 0.589 | 1.2887 | 500 | 0.6300 |
| 0.6007 | 1.3531 | 525 | 0.6289 |
| 0.5751 | 1.4175 | 550 | 0.6284 |
| 0.5627 | 1.4820 | 575 | 0.6275 |
| 0.5689 | 1.5464 | 600 | 0.6267 |
| 0.5098 | 1.6108 | 625 | 0.6259 |
| 0.5623 | 1.6753 | 650 | 0.6248 |
| 0.5608 | 1.7397 | 675 | 0.6237 |
| 0.5552 | 1.8041 | 700 | 0.6233 |
| 0.554 | 1.8686 | 725 | 0.6228 |
| 0.5696 | 1.9330 | 750 | 0.6217 |
| 0.5537 | 1.9974 | 775 | 0.6204 |
### Framework versions
- Transformers 4.49.0
- Pytorch 2.4.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.1