Model Selection - General

Model Configuration

The first configuration block of the Axolotl YAML configuration file is 'model type'. It comprises three main configurations.

1. base_model - This is the Huggingface model that contains *.pt, *.safetensors, or *.bin files

2. model_type - If you want to specify the type of model to load

3. tokenizer_type - Corresponding tokenizer for the model

Model Configuration example for Phi 2.0

base_model: microsoft/phi-2
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer

Model Configuration example for Llama2

base_model: NousResearch/Llama-2-7b-hf
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer
is_llama_derived_model: true

Model Configuration example for TinyLLama

base_model: TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer

Model Configuration example for Code-LLama

base_model: codellama/CodeLlama-7b-hf
model_type: LlamaForCausalLM
tokenizer_type: CodeLlamaTokenizer

We will now move to setting up the Axolotl training configuration file for Phi 2.0.