ahmedsamirio/mistral-sql-create-context-lora

See axolotl config

axolotl version: 0.4.1

base_model: mistralai/Mistral-7B-v0.3
model_type: MistralForCausalLM
tokenizer_type: LlamaTokenizer

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: b-mc2/sql-create-context
    type:
      # JSONL file contains question, context, answer fields per line.
      # This gets mapped to instruction, input, output axolotl tags.
      field_instruction: question
      field_input: context
      field_output: answer
      # Format is used by axolotl to generate the prompt.
      format: |-
        [INST] Using the schema context below, generate a SQL query that answers the question.
        {input}
        {instruction} [/INST]

tokens: # add new control tokens from the dataset to the model
  - "[INST]"
  - " [/INST]"
  - "[SQL]"
  - " [/SQL]"
    
dataset_prepared_path:
val_set_size: 0.05
output_dir: ./outputs/mistral-sql-create-context-lora
hub_model_id: ahmedsamirio/mistral-sql-create-context-lora

# This is set to 4096 in the modal config, why?
# Since I'm using sample packing, decreasing the sequence length will create smaller batches
# which can fit better into memory
sequence_len: 8192

# These is set to false in the modal example, why? (Modal also uses FSDP which might be a reason)
sample_packing: true
eval_sample_packing: true
pad_to_sequence_len: true

adapter: lora
lora_model_dir:
lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:

lora_modules_to_save: # required when adding new tokens to LLaMA/Mistral
  - embed_tokens
  - lm_head
  
lora_target_modules:
  - gate_proj
  - down_proj
  - up_proj
  - q_proj
  - v_proj
  - k_proj
  - o_proj

wandb_project: mistral-sql-create-context
wandb_entity: ahmedsamirio
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 2
micro_batch_size: 4
num_epochs: 1
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

# What is this?
loss_watchdog_threshold: 5.0
loss_watchdog_patience: 3

warmup_steps: 10
evals_per_epoch: 4
eval_table_size:

# This wasn't set in modal config

eval_max_new_tokens: 128
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
  bos_token: "<s>"
  eos_token: "</s>"
  unk_token: "<unk>"