chore: import upstream snapshot with attribution

examples/models/qwen3_next/mcore.sh

@@ -0,0 +1,58 @@
+# 8 * 56GiB
+# padding_free/packing support requires "mcore-bridge>=1.4.0":
+# https://github.com/modelscope/mcore-bridge/pull/76
+PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+NPROC_PER_NODE=8 \
+CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
+megatron sft \
+    --model Qwen/Qwen3-Next-80B-A3B-Instruct \
+    --save_safetensors true \
+    --merge_lora false \
+    --dataset 'swift/Chinese-Qwen3-235B-2507-Distill-data-110k-SFT#2000' \
+              'swift/self-cognition#1000' \
+    --load_from_cache_file true \
+    --tuner_type lora \
+    --lora_rank 8 \
+    --lora_alpha 32 \
+    --packing true \
+    --target_modules all-linear \
+    --tensor_model_parallel_size 2 \
+    --expert_model_parallel_size 4 \
+    --moe_permute_fusion true \
+    --moe_grouped_gemm true \
+    --moe_shared_expert_overlap true \
+    --moe_aux_loss_coeff 1e-6 \
+    --micro_batch_size 1 \
+    --global_batch_size 4 \
+    --recompute_granularity full \
+    --recompute_method uniform \
+    --recompute_num_layers 1 \
+    --num_train_epochs 1 \
+    --finetune true \
+    --cross_entropy_loss_fusion true \
+    --lr 1e-4 \
+    --lr_warmup_fraction 0.05 \
+    --min_lr 1e-5 \
+    --output_dir megatron_output/Qwen3-Next-80B-A3B-Instruct \
+    --eval_steps 200 \
+    --save_steps 200 \
+    --max_length 4096 \
+    --dataloader_num_workers 8 \
+    --dataset_num_proc 8 \
+    --no_save_optim true \
+    --no_save_rng true \
+    --sequence_parallel true \
+    --attention_backend flash \
+    --model_author swift \
+    --model_name swift-robot
+
+
+# CUDA_VISIBLE_DEVICES=0,1,2,3 \
+# swift infer \
+#     --adapters megatron_output/Qwen3-Next-80B-A3B-Instruct/vx-xxx/checkpoint-xxx \
+#     --vllm_tensor_parallel_size 4 \
+#     --infer_backend vllm \
+#     --vllm_max_model_len 8192 \
+#     --vllm_gpu_memory_utilization 0.9 \
+#     --stream true \
+#     --max_new_tokens 2048

examples/models/qwen3_next/mtp.sh

@@ -0,0 +1,55 @@
+# 8 * 60GiB
+
+PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+NPROC_PER_NODE=8 \
+CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
+megatron sft \
+    --model Qwen/Qwen3-Next-80B-A3B-Instruct \
+    --save_safetensors true \
+    --mtp_num_layers 1 \
+    --dataset 'swift/Chinese-Qwen3-235B-2507-Distill-data-110k-SFT#2000' \
+              'swift/self-cognition#1000' \
+    --load_from_cache_file true \
+    --tuner_type lora \
+    --lora_rank 8 \
+    --lora_alpha 32 \
+    --target_modules all-linear \
+    --expert_model_parallel_size 4 \
+    --moe_permute_fusion true \
+    --moe_grouped_gemm true \
+    --moe_shared_expert_overlap true \
+    --moe_aux_loss_coeff 1e-6 \
+    --micro_batch_size 2 \
+    --global_batch_size 16 \
+    --recompute_granularity full \
+    --recompute_method uniform \
+    --recompute_num_layers 1 \
+    --num_train_epochs 1 \
+    --finetune true \
+    --cross_entropy_loss_fusion true \
+    --lr 1e-4 \
+    --lr_warmup_fraction 0.05 \
+    --min_lr 1e-5 \
+    --output_dir megatron_output/Qwen3-Next-80B-A3B-Instruct \
+    --eval_steps 200 \
+    --save_steps 200 \
+    --max_length 2048 \
+    --dataloader_num_workers 8 \
+    --dataset_num_proc 8 \
+    --no_save_optim true \
+    --no_save_rng true \
+    --sequence_parallel true \
+    --attention_backend flash \
+    --model_author swift \
+    --model_name swift-robot
+
+
+# CUDA_VISIBLE_DEVICES=0,1,2,3 \
+# swift infer \
+#     --model megatron_output/Qwen3-Next-80B-A3B-Instruct/vx-xxx/checkpoint-xxx-merged \
+#     --vllm_tensor_parallel_size 4 \
+#     --infer_backend vllm \
+#     --vllm_max_model_len 8192 \
+#     --vllm_gpu_memory_utilization 0.9 \
+#     --vllm_speculative_config '{"method":"qwen3_next_mtp","num_speculative_tokens":2}' \
+#     --max_new_tokens 2048

examples/models/qwen3_next/transformers.sh

@@ -0,0 +1,28 @@
+# 4 * 60GiB
+# Note: Due to linear attention, this model currently does not support padding_free and packing.
+CUDA_VISIBLE_DEVICES=0,1,2,3 \
+swift sft \
+    --model Qwen/Qwen3-Next-80B-A3B-Instruct \
+    --tuner_type lora \
+    --dataset 'swift/self-cognition#1000' \
+    --torch_dtype bfloat16 \
+    --num_train_epochs 1 \
+    --per_device_train_batch_size 1 \
+    --per_device_eval_batch_size 1 \
+    --learning_rate 1e-4 \
+    --lora_rank 8 \
+    --lora_alpha 32 \
+    --target_modules all-linear \
+    --experts_impl grouped_mm \
+    --router_aux_loss_coef 1e-3 \
+    --gradient_accumulation_steps 2 \
+    --eval_steps 50 \
+    --save_steps 50 \
+    --save_total_limit 2 \
+    --logging_steps 5 \
+    --max_length 2048 \
+    --output_dir output \
+    --warmup_ratio 0.05 \
+    --dataloader_num_workers 4 \
+    --model_author swift \
+    --model_name swift-robot