Skip to content

Commit

Permalink
Phi examples (#569)
Browse files Browse the repository at this point in the history
* add phi full ft example

* Add readme to point out that deepspeed should be used

* zero1 is better than zero2 for phi
  • Loading branch information
winglian committed Sep 14, 2023
1 parent c6d870b commit 2284209
Show file tree
Hide file tree
Showing 2 changed files with 82 additions and 0 deletions.
7 changes: 7 additions & 0 deletions examples/phi/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
# Phi

Due to some nuances with the phi code, please use deepspeed when training phi.

```shell
accelerate launch scripts/finetune.py examples/phi/phi-ft.yml --deepspeed deepspeed/zero1.json
```
75 changes: 75 additions & 0 deletions examples/phi/phi-ft.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,75 @@
base_model: microsoft/phi-1_5
base_model_config: microsoft/phi-1_5
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
is_llama_derived_model: false
trust_remote_code: true

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
- path: garage-bAInd/Open-Platypus
type: alpaca

dataset_prepared_path: last_run_prepared
val_set_size: 0.05
output_dir: ./phi-sft-out

sequence_len: 2048
sample_packing: false # does not work with phi
pad_to_sequence_len:

adapter:
lora_model_dir:
lora_r:
lora_alpha:
lora_dropout:
lora_target_linear:
lora_fan_in_fan_out:

wandb_project:
wandb_entity:
wandb_watch:
wandb_run_id:
wandb_log_model:

gradient_accumulation_steps: 2
micro_batch_size: 1
num_epochs: 4
optimizer: adamw_bnb_8bit
adam_beta2: 0.95
adam_epsilon: 0.00001
max_grad_norm: 1.0
lr_scheduler: cosine
learning_rate: 0.000003

train_on_inputs: false
group_by_length: true
bf16: true
fp16: false
tf32: true

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

warmup_steps: 100
eval_steps: 0.05
save_steps:
debug:
deepspeed:
weight_decay: 0.1
fsdp:
fsdp_config:
resize_token_embeddings_to_32x: true
special_tokens:
bos_token: "<|endoftext|>"
eos_token: "<|endoftext|>"
unk_token: "<|endoftext|>"
pad_token: "<|endoftext|>"

0 comments on commit 2284209

Please sign in to comment.