|
|
--- |
|
|
library_name: transformers |
|
|
license: apache-2.0 |
|
|
base_model: Qwen/Qwen3-8B-Base |
|
|
tags: |
|
|
- roleplay |
|
|
- conversational |
|
|
- axolotl |
|
|
- qwen |
|
|
--- |
|
|
|
|
|
# Remnant Qwen3 8b (series 1) |
|
|
|
|
|
[English](./README.md) | [简体中文](./README-cn.md) |
|
|
|
|
|
*There's a wisp of dust in the air. It feels like its from a bygone era, but you don't know where from. It lands on your tongue. It tastes nice.* |
|
|
|
|
|
 |
|
|
|
|
|
Remnant is a series of finetuned LLMs focused on SFW and NSFW roleplaying and conversation. |
|
|
|
|
|
## Quants |
|
|
GGUF: |
|
|
- Todo! |
|
|
|
|
|
EXL3: |
|
|
- Todo! |
|
|
|
|
|
EXL2: |
|
|
- Todo! |
|
|
|
|
|
MISC: |
|
|
- Todo! |
|
|
|
|
|
## Recommended Settings |
|
|
Chat template: ChatML. Apparently Llama 3 format works too, though? Ymmv :3 |
|
|
Samplers: |
|
|
- `0.8` temp |
|
|
- `0.1` min_p |
|
|
- `0.5` presence penalty |
|
|
|
|
|
## Credits |
|
|
Humongous thanks to Allura, ilya <3 |
|
|
Big thanks to the developers of Axolotl (whose training framework I used), Tongyi Qianwen/Qwen/Alibaba (whose model I used), Prime Intellect (whose GPUs I used), and my bank (whose debit card I used) |
|
|
|
|
|
## Misc |
|
|
|
|
|
[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) |
|
|
<details><summary>See axolotl config</summary> |
|
|
|
|
|
axolotl version: `0.10.0.dev0` |
|
|
```yaml |
|
|
# === Model Configuration === |
|
|
base_model: Qwen/Qwen3-8B-Base |
|
|
load_in_8bit: false |
|
|
load_in_4bit: false |
|
|
|
|
|
# === Training Setup === |
|
|
num_epochs: 2 |
|
|
micro_batch_size: 32 |
|
|
gradient_accumulation_steps: 1 |
|
|
sequence_len: 8192 |
|
|
sample_packing: true |
|
|
pad_to_sequence_len: true |
|
|
|
|
|
# === Hyperparameter Configuration === |
|
|
optimizer: apollo_adamw_layerwise |
|
|
# Apollo-mini configuration: |
|
|
optim_args: "proj=random,rank=1,scale=128.0,scale_type=tensor,update_proj_gap=200" |
|
|
# Regular Apollo configuration: |
|
|
# optim_args: |
|
|
optim_target_modules: all_linear |
|
|
learning_rate: 2e-5 |
|
|
lr_scheduler: rex |
|
|
weight_decay: 0.01 |
|
|
warmup_ratio: 0 |
|
|
|
|
|
# === Data Configuration === |
|
|
datasets: |
|
|
- path: allura-org/inkmix-v3.0 |
|
|
type: chat_template |
|
|
split: train |
|
|
field_messages: conversations |
|
|
message_field_role: from |
|
|
message_field_content: value |
|
|
|
|
|
dataset_prepared_path: last_run_prepared |
|
|
chat_template: chatml |
|
|
|
|
|
# === Plugins === |
|
|
plugins: |
|
|
- axolotl.integrations.liger.LigerPlugin |
|
|
- axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin |
|
|
|
|
|
# === Hardware Optimization === |
|
|
gradient_checkpointing: unsloth |
|
|
gradient_checkpointing_kwargs: |
|
|
use_reentrant: false |
|
|
liger_rope: true |
|
|
liger_rms_norm: true |
|
|
liger_glu_activation: true |
|
|
cut_cross_entropy: true |
|
|
|
|
|
# === Wandb Tracking === |
|
|
wandb_project: qwen3-8b-inkmix-v3 |
|
|
|
|
|
# === Checkpointing === |
|
|
saves_per_epoch: 2 |
|
|
save_total_limit: 3 |
|
|
|
|
|
# === Advanced Settings === |
|
|
output_dir: /ephemeral/ckpts |
|
|
bf16: auto |
|
|
flash_attention: true |
|
|
train_on_inputs: false |
|
|
group_by_length: false |
|
|
logging_steps: 1 |
|
|
trust_remote_code: true |
|
|
|
|
|
``` |
|
|
|
|
|
</details> |
