poisson-fish/Causal-Dolphin-Agent-LoRA

Causal-Dolphin-Agent-v0.1

This model is a LoRA fine-tune of CausalLM/7B on Eric's wonderful Dolphin dataset, with THUDM/AgentInstruct mixed in both training runs.

Causal-Dolphin-Agent was trained for 3 epochs on 5 million GPT3.5 augmented FLAN instructions & AgentInstruct dataset in ChatML format. It was then trained a further 3 epochs on 1 million GPT4 augmented FLAN instructions with AgentInstruct shuffled in as well.

It achieves the following results on the evaluation set:

• Loss: 2.8435

Prompt Format

Causal-Dolphin-Agent uses ChatML as the prompt format:

<|im_start|>system
You are Dolphin, a helpful AI assistant.<|im_end|>
<|im_start|>user
If Danny owns a bike, then Edward owns a bike. If Edward owns a bike, then Freddy owns a bike. If Danny owns a bike, which of the following statements must be true? Let's think step by step.

I. Edward owns a bike.
II. Freddy owns a bike.
III. Freddy does not own a bike.

Choose one answer:
I only
II only
III only
I and II only
I and III only
<|im_end|>
<|im_start|>assistant

Training and evaluation data

ehartford/dolphin THUDM/AgentInstruct

Training procedure

Causal-Dolphin-Agent was trained for 3 epochs on 5 million GPT3.5 augmented FLAN instructions & AgentInstruct dataset in ChatML format. It was then trained a further 3 epochs on 1 million GPT4 augmented FLAN instructions with AgentInstruct shuffled in as well.

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 6e-06
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-05
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
2.7774	0.0	1	5.1009
3.2798	0.15	46	5.1010
2.0722	0.3	92	5.0489
2.5919	0.45	138	4.8834
2.0011	0.6	184	4.6678
1.3733	0.75	230	4.4628
1.7321	0.9	276	4.2757
1.3994	1.05	322	4.1029
1.2308	1.2	368	3.8916
0.8229	1.35	414	3.6451
0.9592	1.5	460	3.4106
0.8528	1.65	506	3.2250
0.7362	1.8	552	3.0852
0.8077	1.95	598	2.9881
0.6912	2.1	644	2.9315
0.7776	2.25	690	2.8911
0.6916	2.41	736	2.8678
0.8674	2.56	782	2.8534
0.7797	2.71	828	2.8545
0.6838	2.86	874	2.8435

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0+cu121
• Datasets 2.14.6
• Tokenizers 0.14.1