Hugging Face's logo

Gaie
/
alpaca-2-7b-reproduced

Text Generation
Transformers
Safetensors
llama
text-generation-inference
Inference Endpoints
Model card Files Community
alpaca-2-7b-reproduced
File size: 3,257 Bytes 
502750c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
 

#!/usr/bin/env bash
#
# Copyright 2023 PKU-Alignment Team. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================

if [ -z "${BASH_VERSION}" ]; then
	echo "Please use bash to run this script." >&2
	exit 1
fi

set -x

SCRIPT_DIR="$(cd "$(dirname "$0")" &>/dev/null && pwd)"
ROOT_DIR="$(dirname "${SCRIPT_DIR}")"
export PYTHONPATH="${ROOT_DIR}${PYTHONPATH:+:${PYTHONPATH}}"
export LOGLEVEL="${LOGLEVEL:-WARNING}"

MODEL_NAME_OR_PATH="/mnt/data/jiayi/Llama-2-7b-hf"
OUTPUT_DIR="${ROOT_DIR}/alpaca-2"
unset HOSTFILE
ZERO_STAGE=3
OFFLOAD="none"
while [[ "$#" -gt 0 ]]; do
	arg="$1"
	shift
	case "${arg}" in
		--model_name_or_path)
			MODEL_NAME_OR_PATH="$1"
			shift
			;;
		--model_name_or_path=*)
			MODEL_NAME_OR_PATH="${arg#*=}"
			;;
		--output_dir)
			OUTPUT_DIR="$1"
			shift
			;;
		--output_dir=*)
			OUTPUT_DIR="${arg#*=}"
			;;
		--hostfile)
			HOSTFILE="$1"
			shift
			;;
		--hostfile=*)
			HOSTFILE="${arg#*=}"
			;;
		--zero_stage)
			ZERO_STAGE="$1"
			shift
			;;
		--zero_stage=*)
			ZERO_STAGE="${arg#*=}"
			;;
		--offload)
			OFFLOAD="$1"
			shift
			;;
		--offload=*)
			OFFLOAD="${arg#*=}"
			;;
		*)
			echo "Unknown parameter passed: '${arg}'" >&2
			exit 1
			;;
	esac
done

mkdir -p "${OUTPUT_DIR}"
OUTPUT_DIR="$(cd "${OUTPUT_DIR}" &>/dev/null && pwd)"
if [[ ! -f "${OUTPUT_DIR}/.gitignore" ]]; then
	echo '*' >"${OUTPUT_DIR}/.gitignore"
fi

cp -f "$0" "${OUTPUT_DIR}/script.sh"

if [[ -z "${WANDB_API_KEY}" ]]; then
	export WANDB_MODE="offline"
fi

MASTER_PORT_START=10000
MASTER_PORT_END=65535
MASTER_PORT="$(
	comm -23 \
		<(seq "${MASTER_PORT_START}" "${MASTER_PORT_END}" | sort) \
		<(ss -Htan | awk '{ print $4 }' | awk -F ':' '{ print $NF }' | sort -u) |
		shuf | head -n 1
)"

DEEPSPEED_ARGS=()
if [[ -n "${HOSTFILE+x}" ]]; then
	DEEPSPEED_ARGS+=("--hostfile" "${HOSTFILE}")
fi
DEEPSPEED_ARGS+=("--master_port" "${MASTER_PORT}")

exec 1> >(tee "${OUTPUT_DIR}/stdout.log" >&1) 2> >(tee "${OUTPUT_DIR}/stderr.log" >&2)

torchrun --nproc_per_node=8 --master_port="${MASTER_PORT}" train.py \
    --model_name_or_path "${MODEL_NAME_OR_PATH}" \
    --data_path ./alpaca_data.json \
    --bf16 True \
    --output_dir "${OUTPUT_DIR}" \
    --num_train_epochs 3 \
    --per_device_train_batch_size 4 \
    --per_device_eval_batch_size 4 \
    --gradient_accumulation_steps 8 \
    --evaluation_strategy "no" \
    --save_strategy "steps" \
    --save_steps 2000 \
    --save_total_limit 1 \
    --learning_rate 2e-5 \
    --weight_decay 0. \
    --warmup_ratio 0.03 \
    --lr_scheduler_type "cosine" \
    --logging_steps 1 \
    --fsdp "full_shard auto_wrap" \
    --fsdp_transformer_layer_cls_to_wrap 'LlamaDecoderLayer' \
    --tf32 True