chore: import upstream snapshot with attribution

examples/train/embedding/qwen3/infer.py

@@ -0,0 +1,45 @@
+# For full-parameter training, please refer to:
+# https://github.com/modelscope/ms-swift/blob/main/examples/infer/demo_embedding.py
+
+import torch
+
+from swift.infer_engine import InferRequest, TransformersEngine
+
+
+def run_qwen3_emb():
+    engine = TransformersEngine(
+        'Qwen/Qwen3-Embedding-4B',
+        task_type='embedding',
+        attn_impl='flash_attention_2',
+        adapters=['output/vx-xxx/checkpoint-xxx'])
+
+    infer_requests = [
+        InferRequest(messages=[
+            {
+                'role': 'user',
+                'content': 'A dog sleeping under a table.'
+            },
+        ]),
+        InferRequest(messages=[
+            {
+                'role': 'user',
+                'content': 'a dog napping under a small table.'
+            },
+        ]),
+        InferRequest(messages=[
+            {
+                'role': 'user',
+                'content': 'a cat napping under a small tree.'
+            },
+        ])
+    ]
+    resp_list = engine.infer(infer_requests)
+    embedding0 = torch.tensor(resp_list[0].data[0].embedding)
+    embedding1 = torch.tensor(resp_list[1].data[0].embedding)
+    embedding2 = torch.tensor(resp_list[2].data[0].embedding)
+    embedding = torch.stack([embedding0, embedding1, embedding2])
+    print(f'scores: {embedding @ embedding.T}')
+
+
+if __name__ == '__main__':
+    run_qwen3_emb()

examples/train/embedding/qwen3/qwen3_emb.sh

@@ -0,0 +1,39 @@
+# 2*10GiB
+# losses: swift/loss
+# data format: docs/source_en/BestPractices/Embedding.md
+# --dataloader_drop_last must be true or eval gather will throw error
+# --model iic/gte-modernbert-base iic/gte_Qwen2-7B-instruct also supported
+CUDA_VISIBLE_DEVICES=0,1 \
+INFONCE_TEMPERATURE=0.1 \
+NPROC_PER_NODE=2 \
+swift sft \
+    --model Qwen/Qwen3-Embedding-4B \
+    --task_type embedding \
+    --tuner_type lora \
+    --lora_rank 8 \
+    --lora_alpha 32 \
+    --learning_rate 5e-5 \
+    --target_modules all-linear \
+    --dataset sentence-transformers/stsb:positive \
+    --attn_impl flash_attn \
+    --padding_free true \
+    --torch_dtype bfloat16 \
+    --load_from_cache_file true \
+    --split_dataset_ratio 0.02 \
+    --eval_strategy steps \
+    --output_dir output \
+    --save_steps 50 \
+    --eval_steps 50 \
+    --save_total_limit 2 \
+    --logging_steps 5 \
+    --num_train_epochs 5 \
+    --max_length 8192 \
+    --per_device_train_batch_size 8 \
+    --per_device_eval_batch_size 8 \
+    --gradient_accumulation_steps 1 \
+    --dataloader_num_workers 4 \
+    --dataset_num_proc 4 \
+    --warmup_ratio 0.05 \
+    --loss_type infonce \
+    --dataloader_drop_last true \
+    --deepspeed zero2

examples/train/embedding/qwen3/qwen3_vl_emb.sh

@@ -0,0 +1,35 @@
+# 2*30GiB
+CUDA_VISIBLE_DEVICES=0,1 \
+INFONCE_TEMPERATURE=0.1 \
+NPROC_PER_NODE=2 \
+swift sft \
+    --model Qwen/Qwen3-VL-Embedding-8B \
+    --task_type embedding \
+    --tuner_type lora \
+    --lora_rank 8 \
+    --lora_alpha 32 \
+    --learning_rate 5e-5 \
+    --target_modules all-linear \
+    --dataset swift/TextCaps:emb \
+    --attn_impl flash_attn \
+    --padding_free true \
+    --torch_dtype bfloat16 \
+    --load_from_cache_file true \
+    --split_dataset_ratio 0.02 \
+    --eval_strategy steps \
+    --output_dir output \
+    --save_steps 50 \
+    --eval_steps 50 \
+    --save_total_limit 2 \
+    --logging_steps 5 \
+    --num_train_epochs 1 \
+    --max_length 8192 \
+    --per_device_train_batch_size 8 \
+    --per_device_eval_batch_size 8 \
+    --gradient_accumulation_steps 1 \
+    --dataloader_num_workers 4 \
+    --dataset_num_proc 4 \
+    --warmup_ratio 0.05 \
+    --loss_type infonce \
+    --dataloader_drop_last true \
+    --deepspeed zero2

examples/train/embedding/train_gme.sh

@@ -0,0 +1,31 @@
+nproc_per_node=8
+
+# losses: swift/loss
+# 8*40G
+MAX_PIXELS=1003520 \
+NPROC_PER_NODE=$nproc_per_node \
+swift sft \
+    --model iic/gme-Qwen2-VL-2B-Instruct \
+    --tuner_type lora \
+    --dataset 'swift/TextCaps:emb' \
+    --load_from_cache_file true \
+    --split_dataset_ratio 0.01 \
+    --torch_dtype bfloat16 \
+    --num_train_epochs 1 \
+    --per_device_train_batch_size 2 \
+    --per_device_eval_batch_size 2 \
+    --gradient_accumulation_steps $(expr 64 / $nproc_per_node) \
+    --eval_steps 100 \
+    --save_steps 100 \
+    --eval_strategy steps \
+    --save_total_limit 2 \
+    --logging_steps 5 \
+    --output_dir output \
+    --lazy_tokenize true \
+    --warmup_ratio 0.05 \
+    --learning_rate 5e-5 \
+    --deepspeed zero3 \
+    --dataloader_num_workers 4 \
+    --task_type embedding \
+    --loss_type infonce \
+    --dataloader_drop_last true