Delta-Vector/SD-prompter-4B-V0.2-exl2

exl2 quant (measurement.json in main branch)

---

check revisions for quants

---

🔍 Overview

This is the second in a line of models dedicated to creating Stable-Diffusion prompts when given a character appearance. Made for the CharGen Project, This has been finetuned ontop of Delta-Vector/Holland-4B-V1

⚖️ Quants

Available quantization formats:

• GGUF: https://huggingface.co/mradermacher/SDPrompter4b-GGUF
• EXL2: https://huggingface.co/Delta-Vector/SD-prompter-4B-V0.2-exl2/

💬 Prompting

Recommended format: ChatML, Use the following system prompt for the model. I'd advise against setting a high amount of output tokens as the model loops, use 0.1 min-p and temp-1 to keep it coherent.

Create a prompt for Stable Diffusion based on the information below.

🌟 Credits

Finetuned on 1xRTX6000 provided by Kubernetes_bad, All credits goes to Kubernetes_bad, LucyKnada and the rest of Anthracite.

🛠️ Axolotl Config)

base_model: Delta-Vector/Holland-4B-V1
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
trust_remote_code: true
load_in_8bit: false
load_in_4bit: false
strict: false
datasets:
  - path: NewEden/CivitAI-SD-Prompts
datasets:
  - path: NewEden/CivitAI-Prompts-Sharegpt
    type: chat_template
    chat_template: chatml
    roles_to_train: ["gpt"]
    field_messages: conversations
    message_field_role: from
    message_field_content: value
    train_on_eos: turn
dataset_prepared_path:
val_set_size: 0.02
output_dir: ./outputs/out2
sequence_len: 8192
sample_packing: true
eval_sample_packing: false
pad_to_sequence_len: true
plugins:
  - axolotl.integrations.liger.LigerPlugin
liger_rope: true
liger_rms_norm: true
liger_swiglu: true
liger_fused_linear_cross_entropy: true
wandb_project: SDprompter-final
wandb_entity:
wandb_watch:
wandb_name: SDprompter-final
wandb_log_model:
gradient_accumulation_steps: 16
micro_batch_size: 1
num_epochs: 4
optimizer: paged_adamw_8bit
lr_scheduler: cosine
learning_rate: 0.00001
train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: true
gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
warmup_ratio: 0.05
evals_per_epoch: 4
saves_per_epoch: 1
debug:
weight_decay: 0.01
special_tokens:
  pad_token: <|finetune_right_pad_id|>
  eos_token: <|eot_id|>
auto_resume_from_checkpoints: true