[AutoParallel] Add benchmark for llama-7b-dy2st.

tests/test_tipc/static/auto_parallel/llama2/N1C8/meta-llama-Llama-2-7b_pretrain_dy2st_bs8_Sharding8_Stage1.sh

@@ -0,0 +1,25 @@
+# Copyright (c) 2024 PaddlePaddle Authors. 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.
+
+param="model_item=meta-llama-Llama-2-7b_pretrain_dy2st "
+param+="run_mode=Sharding_Stage1 "
+param+="device_num=N1C8 "
+param+="global_batch_size=8 "
+param+="nnodes=1 "
+param+="model_type=llama2_7b "
+
+cd ./tests
+bash ./test_tipc/static/auto_parallel/llama2/benchmark_common/prepare.sh
+
+bash -c "${param} bash ./test_tipc/static/auto_parallel/llama2/benchmark_common/run_benchmark.sh"

tests/test_tipc/static/auto_parallel/llama2/pretrain_config_llama2_7b/pretrain-llama2_7b.json

@@ -0,0 +1,52 @@
+{
+  "model_name_or_path": "meta-llama/Llama-2-7b",
+  "tokenizer_name_or_path": "meta-llama/Llama-2-7b",
+  "input_dir": "./data",
+  "output_dir": "./checkpoints/llama2_pretrain_ckpts",
+  "per_device_train_batch_size": 1,
+  "gradient_accumulation_steps": 1,
+  "per_device_eval_batch_size": 2,
+  "tensor_parallel_degree": 1,
+  "pipeline_parallel_degree": 1,
+  "sharding": "stage1",
+  "sharding_parallel_config": "enable_stage1_overlap",
+  "tensor_parallel_config": "enable_delay_scale_loss enable_mp_async_allreduce enable_mp_skip_c_identity enable_mp_fused_linear_param_grad_add",
+  "pipeline_parallel_config": "enable_delay_scale_loss enable_release_grads disable_partial_send_recv",
+  "virtual_pp_degree": 1,
+  "sequence_parallel": 0,   
+  "use_flash_attention": true,
+  "use_fused_rms_norm": true,
+  "fuse_attention_ffn": true,
+  "fuse_attention_qkv": true,
+  "use_fused_rope": true,
+  "fused_linear_param_grad_add": true,
+  "max_seq_length": 4096,
+  "learning_rate": 3e-05,
+  "min_learning_rate": 3e-06,
+  "warmup_steps": 30,
+  "logging_steps": 1,
+  "max_steps": 50,
+  "save_steps": 5000,
+  "eval_steps": 1000,
+  "weight_decay": 0.01,
+  "bf16": true,
+  "fp16_opt_level": "O2",
+  "amp_master_grad": true,
+  "warmup_ratio": 0.01,
+  "max_grad_norm": 1.0,
+  "dataloader_num_workers": 1,
+  "continue_training": 0,
+  "do_train": true,
+  "do_eval": false,
+  "do_predict": false,
+  "disable_tqdm": true,
+  "skip_profile_timer": true,
+  "recompute": false,
+  "recompute_use_reentrant": true,
+  "distributed_dataloader": 0,
+  "recompute_granularity": "full",
+  "save_total_limit": 2,
+  "device": "gpu",
+  "to_static": true,
+  "enable_auto_parallel": true
+}