Felladrin/Pythia-31M-Chat-v1

A Pythia Chat Model of 31M Parameters

• Base model: EleutherAI/pythia-31m
• Availability in other ML formats:
  • GGUF: Felladrin/gguf-Pythia-31M-Chat-v1
  • ONNX: Felladrin/onnx-Pythia-31M-Chat-v1

Recommended prompt format

<|im_start|>system
{system_message}<|im_end|>
<|im_start|>user
{user_message}<|im_end|>
<|im_start|>assistant

Recommended inference parameters

penalty_alpha: 0.5
top_k: 2
repetition_penalty: 1.0016

Datasets and parameters used for training

Dataset	License Type
totally-not-an-llm/EverythingLM-data-V3	mit
databricks/databricks-dolly-15k	cc-by-sa-3.0
THUDM/webglm-qa	apache-2.0
starfishmedical/webGPT_x_dolly	cc-by-sa-3.0
Amod/mental_health_counseling_conversations	openrail
sablo/oasst2_curated	apache-2.0
cognitivecomputations/wizard_vicuna_70k_unfiltered	apache-2.0
mlabonne/chatml_dpo_pairs	apache-2.0

SFTTrainer(
    model,
    train_dataset=train_dataset,
    dataset_text_field="text",
    eval_dataset=eval_dataset,
    max_seq_length=2048,
    packing=True,
    args=TrainingArguments(
        learning_rate=2e-6,
        per_device_train_batch_size=1,
        per_device_eval_batch_size=1,
        gradient_accumulation_steps=16,
        lr_scheduler_type="cosine",
        num_train_epochs=1,
        logging_strategy="steps",
        save_strategy="steps",
        evaluation_strategy="steps",
        logging_steps=10,
        eval_steps=10,
        save_steps=10,
        warmup_steps=50,
        load_best_model_at_end=True,
        metric_for_best_model="eval_loss",
        greater_is_better=False,
        weight_decay=0.01,
        save_total_limit=10,
        neftune_noise_alpha=5,
    ),
    callbacks=[
        EarlyStoppingCallback(
            early_stopping_patience=3,
            early_stopping_threshold=0.005
        ),
    ],
)

DPOTrainer(
    model,
    beta=0.1,
    train_dataset=dataset,
    tokenizer=tokenizer,
    eval_dataset=eval_dataset,
    max_length=1536,
    max_prompt_length=1024,
    args=TrainingArguments(
        learning_rate=2e-6,
        per_device_train_batch_size=1,
        per_device_eval_batch_size=1,
        gradient_accumulation_steps=1,
        lr_scheduler_type="cosine",
        num_train_epochs=1,
        logging_strategy="steps",
        save_strategy="steps",
        evaluation_strategy="steps",
        logging_steps=1,
        eval_steps=1,
        save_steps=1,
        warmup_steps=0,
        load_best_model_at_end=True,
        metric_for_best_model="eval_loss",
        greater_is_better=False,
        weight_decay=0.0,
        neftune_noise_alpha=5,
        remove_unused_columns=False,
    ),
    callbacks=[
        EarlyStoppingCallback(
            early_stopping_patience=3,
            early_stopping_threshold=0.005
        ),
    ],
)

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	19.92
AI2 Reasoning Challenge (25-Shot)	22.70
HellaSwag (10-Shot)	25.60
MMLU (5-Shot)	23.24
TruthfulQA (0-shot)	0.00
Winogrande (5-shot)	47.99
GSM8k (5-shot)	0.00