chore: import upstream snapshot with attribution

examples/models/qwen3_5/fp8.sh

@@ -0,0 +1,81 @@
+# 2 * 50GiB
+# `--model` can directly accept a model with FP8 weights,
+# and the FP8 weight model will be saved directly upon training completion.
+PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+NPROC_PER_NODE=2 \
+CUDA_VISIBLE_DEVICES=0,1 \
+IMAGE_MAX_TOKEN_NUM=1024 \
+VIDEO_MAX_TOKEN_NUM=128 \
+FPS_MAX_FRAMES=12 \
+megatron sft \
+    --model Qwen/Qwen3.5-4B \
+    --save_safetensors true \
+    --dataset 'AI-ModelScope/alpaca-gpt4-data-zh#500' \
+              'AI-ModelScope/alpaca-gpt4-data-en#500' \
+              'swift/self-cognition#500' \
+              'AI-ModelScope/LaTeX_OCR:human_handwrite#2000' \
+    --model_author swift \
+    --model_name swift-robot \
+    --linear_decoupled_in_proj true \
+    --load_from_cache_file true \
+    --add_non_thinking_prefix true \
+    --loss_scale ignore_empty_think \
+    --fp8_recipe blockwise \
+    --fp8_format e4m3 \
+    --fp8_param_gather true \
+    --split_dataset_ratio 0.01 \
+    --tuner_type full \
+    --tensor_model_parallel_size 2 \
+    --micro_batch_size 1 \
+    --global_batch_size 2 \
+    --recompute_granularity full \
+    --recompute_method uniform \
+    --recompute_num_layers 1 \
+    --num_train_epochs 1 \
+    --packing true \
+    --finetune true \
+    --freeze_llm false \
+    --freeze_vit true \
+    --freeze_aligner true \
+    --cross_entropy_loss_fusion true \
+    --lr 1e-5 \
+    --lr_warmup_fraction 0.05 \
+    --min_lr 1e-6 \
+    --output_dir megatron_output/Qwen3.5-4B-FP8 \
+    --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 \
+    --mtp_num_layers 1 \
+    --attention_backend flash
+
+# PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+# CUDA_VISIBLE_DEVICES=0 \
+# IMAGE_MAX_TOKEN_NUM=1024 \
+# VIDEO_MAX_TOKEN_NUM=128 \
+# FPS_MAX_FRAMES=12 \
+# swift infer \
+#     --model megatron_output/Qwen3.5-4B-FP8/vx-xxx/checkpoint-xxx \
+#     --stream true \
+#     --enable_thinking false \
+#     --load_data_args true
+
+# vllm
+# PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+# CUDA_VISIBLE_DEVICES=0 \
+# IMAGE_MAX_TOKEN_NUM=1024 \
+# VIDEO_MAX_TOKEN_NUM=128 \
+# FPS_MAX_FRAMES=12 \
+# swift infer \
+#     --model megatron_output/Qwen3.5-4B-FP8/vx-xxx/checkpoint-xxx \
+#     --stream true \
+#     --enable_thinking false \
+#     --load_data_args true \
+#     --infer_backend vllm \
+#     --vllm_tensor_parallel_size 1 \
+#     --vllm_max_model_len 8192 \
+#     --vllm_speculative_config '{"method":"qwen3_next_mtp","num_speculative_tokens":2}'

examples/models/qwen3_5/mcore.sh

@@ -0,0 +1,70 @@
+# 4 * 40GiB
+
+PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+NPROC_PER_NODE=4 \
+CUDA_VISIBLE_DEVICES=0,1,2,3 \
+IMAGE_MAX_TOKEN_NUM=1024 \
+VIDEO_MAX_TOKEN_NUM=128 \
+FPS_MAX_FRAMES=12 \
+megatron sft \
+    --model Qwen/Qwen3.5-35B-A3B \
+    --save_safetensors true \
+    --merge_lora true \
+    --dataset 'AI-ModelScope/alpaca-gpt4-data-zh#500' \
+              'AI-ModelScope/alpaca-gpt4-data-en#500' \
+              'swift/self-cognition#500' \
+              'AI-ModelScope/LaTeX_OCR:human_handwrite#2000' \
+    --load_from_cache_file true \
+    --add_non_thinking_prefix true \
+    --loss_scale ignore_empty_think \
+    --split_dataset_ratio 0.01 \
+    --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 4 \
+    --global_batch_size 16 \
+    --recompute_granularity full \
+    --recompute_method uniform \
+    --recompute_num_layers 1 \
+    --num_train_epochs 1 \
+    --group_by_length true \
+    --finetune true \
+    --freeze_llm false \
+    --freeze_vit true \
+    --freeze_aligner true \
+    --cross_entropy_loss_fusion true \
+    --lr 1e-4 \
+    --lr_warmup_fraction 0.05 \
+    --min_lr 1e-5 \
+    --output_dir megatron_output/Qwen3.5-35B-A3B \
+    --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 \
+    --padding_free false \
+    --model_author swift \
+    --model_name swift-robot
+
+
+# PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+# CUDA_VISIBLE_DEVICES=0,1,2,3 \
+# IMAGE_MAX_TOKEN_NUM=1024 \
+# VIDEO_MAX_TOKEN_NUM=128 \
+# FPS_MAX_FRAMES=12 \
+# swift infer \
+#     --model megatron_output/Qwen3.5-35B-A3B/vx-xxx/checkpoint-xxx-merged \
+#     --stream true \
+#     --experts_impl grouped_mm \
+#     --enable_thinking false \
+#     --load_data_args true

examples/models/qwen3_5/mcore_grpo_moe.sh

@@ -0,0 +1,64 @@
+# Qwen3.5-35B-A3B (MoE) Megatron GRPO LoRA Training Example
+
+SYSTEM_PROMPT="""You are a helpful math assistant. Solve the problem step by step and put your final answer within \\boxed{}."""
+
+CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
+NPROC_PER_NODE=8 \
+PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+megatron rlhf \
+    --rlhf_type grpo \
+    --model Qwen/Qwen3.5-35B-A3B \
+    --save_safetensors true \
+    --enable_thinking false \
+    --merge_lora true \
+    --context_parallel_size 1 \
+    --tensor_model_parallel_size 1 \
+    --expert_model_parallel_size 8 \
+    --pipeline_model_parallel_size 1 \
+    --moe_permute_fusion true \
+    --dataset open-r1/DAPO-Math-17k-Processed \
+    --system "$SYSTEM_PROMPT" \
+    --num_train_epochs 1 \
+    --global_batch_size 64 \
+    --micro_batch_size 1 \
+    --steps_per_generation 2 \
+    --num_generations 8 \
+    --reward_funcs accuracy \
+    --use_vllm true \
+    --vllm_mode colocate \
+    --vllm_gpu_memory_utilization 0.5 \
+    --vllm_tensor_parallel_size 2 \
+    --vllm_max_model_len 9192 \
+    --max_length 1000 \
+    --max_completion_length 8192 \
+    --tuner_type lora \
+    --target_modules all-linear \
+    --lr 5e-5 \
+    --bf16 true \
+    --beta 0.00 \
+    --epsilon 0.2 \
+    --epsilon_high 0.28 \
+    --dynamic_sample false \
+    --overlong_filter true \
+    --loss_type grpo \
+    --sleep_level 1 \
+    --offload_model true \
+    --offload_bridge false \
+    --offload_optimizer true \
+    --logging_steps 1 \
+    --recompute_granularity full \
+    --recompute_method uniform \
+    --recompute_num_layers 1 \
+    --finetune \
+    --dataloader_num_workers 8 \
+    --dataset_num_proc 8 \
+    --no_save_optim \
+    --no_save_rng \
+    --save_steps 20 \
+    --attention_backend flash \
+    --moe_expert_capacity_factor 2 \
+    --temperature 1.0 \
+    --padding_free false \
+    --sequence_parallel true \
+    --log_completions true \
+    --report_to tensorboard swanlab

examples/models/qwen3_5/packing.sh

@@ -0,0 +1,51 @@
+# 8 * 80GiB
+PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+NPROC_PER_NODE=8 \
+CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
+IMAGE_MAX_TOKEN_NUM=1024 \
+VIDEO_MAX_TOKEN_NUM=128 \
+FPS_MAX_FRAMES=12 \
+megatron sft \
+    --model Qwen/Qwen3.5-35B-A3B \
+    --save_safetensors true \
+    --dataset 'AI-ModelScope/LongAlpaca-12k' \
+    --load_from_cache_file true \
+    --add_non_thinking_prefix true \
+    --split_dataset_ratio 0.01 \
+    --tuner_type full \
+    --tensor_model_parallel_size 4 \
+    --expert_model_parallel_size 8 \
+    --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 \
+    --packing true \
+    --finetune true \
+    --freeze_llm false \
+    --freeze_vit true \
+    --freeze_aligner true \
+    --cross_entropy_loss_fusion true \
+    --lr 1e-5 \
+    --lr_warmup_fraction 0.05 \
+    --min_lr 1e-6 \
+    --output_dir megatron_output/Qwen3.5-35B-A3B \
+    --eval_steps 200 \
+    --save_steps 200 \
+    --max_length 32768 \
+    --dataloader_num_workers 8 \
+    --dataset_num_proc 8 \
+    --no_save_optim true \
+    --no_save_rng true \
+    --sequence_parallel true \
+    --moe_expert_capacity_factor 2 \
+    --mtp_num_layers 1 \
+    --optimizer_cpu_offload true \
+    --use_precision_aware_optimizer true \
+    --optimizer_offload_fraction 0.64 \
+    --attention_backend flash

examples/models/qwen3_5/transformers.sh

@@ -0,0 +1,49 @@
+# 4 * 30GiB
+PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+NPROC_PER_NODE=4 \
+IMAGE_MAX_TOKEN_NUM=1024 \
+VIDEO_MAX_TOKEN_NUM=128 \
+FPS_MAX_FRAMES=12 \
+CUDA_VISIBLE_DEVICES=0,1,2,3 \
+swift sft \
+    --model Qwen/Qwen3.5-35B-A3B \
+    --tuner_type lora \
+    --dataset 'AI-ModelScope/LaTeX_OCR:human_handwrite#2000' \
+    --load_from_cache_file true \
+    --add_non_thinking_prefix true \
+    --split_dataset_ratio 0.01 \
+    --torch_dtype bfloat16 \
+    --num_train_epochs 1 \
+    --per_device_train_batch_size 4 \
+    --per_device_eval_batch_size 4 \
+    --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 1 \
+    --group_by_length true \
+    --output_dir output/Qwen3.5-35B-A3B \
+    --eval_steps 50 \
+    --save_steps 50 \
+    --save_total_limit 2 \
+    --logging_steps 5 \
+    --max_length 2048 \
+    --warmup_ratio 0.05 \
+    --dataset_num_proc 4 \
+    --dataloader_num_workers 4 \
+    --deepspeed zero3
+
+
+# PYTORCH_CUDA_ALLOC_CONF='expandable_segments:True' \
+# CUDA_VISIBLE_DEVICES=0,1,2,3 \
+# IMAGE_MAX_TOKEN_NUM=1024 \
+# VIDEO_MAX_TOKEN_NUM=128 \
+# FPS_MAX_FRAMES=12 \
+# swift infer \
+#     --adapters output/Qwen3.5-35B-A3B/vx-xxx/checkpoint-xxx \
+#     --stream true \
+#     --experts_impl grouped_mm \
+#     --enable_thinking false \
+#     --load_data_args true