add missing files (#631)

Author: tastelikefeet

examples/pytorch/llm/scripts/grok-1/lora_ddp_ds/infer.sh

@@ -0,0 +1,12 @@
+# Experimental environment: 8 * A100
+# Memory cost: 8 * 80G
+PYTHONPATH=../../.. \
+python llm_infer.py \
+    --ckpt_dir output/grok-1/vxx-xxxx-xxxx/checkpoint-xxx \
+    --dtype bf16 \
+    --load_dataset_config true \
+    --max_new_tokens 64 \
+    --do_sample true \
+    --dtype bf16 \
+    --eval_human false \
+    --merge_lora false \

examples/pytorch/llm/scripts/grok-1/lora_ddp_ds/sft.sh

@@ -0,0 +1,38 @@
+# Experimental environment: 8 * A100
+# Memory cost: 8 * 21G
+nproc_per_node=8
+
+PYTHONPATH=../../.. \
+torchrun \
+    --nproc_per_node=$nproc_per_node \
+    --master_port 29500 \
+    llm_sft.py \
+    --model_type grok-1 \
+    --sft_type lora \
+    --tuner_backend swift \
+    --dtype bf16 \
+    --output_dir output \
+    --ddp_backend nccl \
+    --dataset dureader-robust-zh \
+    --train_dataset_sample -1 \
+    --num_train_epochs 1 \
+    --max_length 512 \
+    --check_dataset_strategy warning \
+    --lora_rank 8 \
+    --lora_alpha 32 \
+    --lora_dropout_p 0.05 \
+    --lora_dtype bf16 \
+    --lora_target_modules DEFAULT \
+    --gradient_checkpointing true \
+    --batch_size 2 \
+    --weight_decay 0.1 \
+    --learning_rate 1e-4 \
+    --gradient_accumulation_steps $(expr 16 / $nproc_per_node) \
+    --max_grad_norm 0.5 \
+    --warmup_ratio 0.03 \
+    --eval_steps 100 \
+    --save_steps 100 \
+    --save_total_limit 2 \
+    --logging_steps 10 \
+    --deepspeed_config_path scripts/grok-1/lora_ddp_ds/zero3.json \
+    --save_only_model true \

examples/pytorch/llm/scripts/grok-1/lora_ddp_ds/zero3.json

@@ -0,0 +1,62 @@
+{
+    "fp16": {
+        "enabled": "auto",
+        "loss_scale": 0,
+        "loss_scale_window": 1000,
+        "initial_scale_power": 16,
+        "hysteresis": 2,
+        "min_loss_scale": 1
+    },
+
+    "bf16": {
+        "enabled": "auto"
+    },
+
+    "optimizer": {
+        "type": "AdamW",
+        "params": {
+            "lr": "auto",
+            "betas": "auto",
+            "eps": "auto",
+            "weight_decay": "auto"
+        }
+    },
+
+    "scheduler": {
+        "type": "WarmupDecayLR",
+        "params": {
+            "total_num_steps": "auto",
+            "warmup_min_lr": "auto",
+            "warmup_max_lr": "auto",
+            "warmup_num_steps": "auto"
+        }
+    },
+
+    "zero_optimization": {
+        "stage": 3,
+        "offload_optimizer": {
+            "device": "none",
+            "pin_memory": true
+        },
+        "offload_param": {
+            "device": "cpu",
+            "pin_memory": true
+        },
+        "overlap_comm": true,
+        "contiguous_gradients": true,
+        "sub_group_size": 1e9,
+        "reduce_bucket_size": "auto",
+        "stage3_prefetch_bucket_size": "auto",
+        "stage3_param_persistence_threshold": "auto",
+        "stage3_max_live_parameters": 1e9,
+        "stage3_max_reuse_distance": 1e9,
+        "stage3_gather_16bit_weights_on_model_save": true
+    },
+
+    "gradient_accumulation_steps": "auto",
+    "gradient_clipping": "auto",
+    "steps_per_print": 2000,
+    "train_batch_size": "auto",
+    "train_micro_batch_size_per_gpu": "auto",
+    "wall_clock_breakdown": false
+}