chore: import upstream snapshot with attribution

2026-07-13 13:34:58 +08:00
commit a203934033
1368 changed files with 175001 additions and 0 deletions
@@ -0,0 +1,42 @@
+# ms-swift>=3.11
+swift export \
+    --model Qwen/Qwen2.5-7B \
+    --dataset 'AI-ModelScope/ruozhiba:all' \
+    --dataset_num_proc 64 \
+    --to_cached_dataset true \
+    --truncation_strategy split \
+    --max_length 8192 \
+    --split_dataset_ratio 0.01 \
+    --use_chat_template false \
+    --loss_scale all \
+    --output_dir ./pretrain_cached_dataset
+
+PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+NPROC_PER_NODE=4 \
+CUDA_VISIBLE_DEVICES=0,1,2,3 \
+swift pt \
+    --model Qwen/Qwen2.5-7B \
+    --tuner_type full \
+    --cached_dataset './pretrain_cached_dataset/train' \
+    --cached_val_dataset './pretrain_cached_dataset/val' \
+    --truncation_strategy split \
+    --num_train_epochs 3 \
+    --torch_dtype bfloat16 \
+    --per_device_train_batch_size 1 \
+    --per_device_eval_batch_size 1 \
+    --learning_rate 1e-5 \
+    --gradient_accumulation_steps 4 \
+    --packing true \
+    --eval_steps 200 \
+    --save_steps 200 \
+    --logging_steps 5 \
+    --max_length 8192 \
+    --warmup_ratio 0.05 \
+    --dataloader_num_workers 8 \
+    --dataset_num_proc 8 \
+    --save_total_limit 2 \
+    --save_only_model true \
+    --output_dir output/Qwen2.5-7B \
+    --deepspeed zero3 \
+    --use_liger_kernel true \
+    --attn_impl flash_attn