github
/
minimind
mirror of https://github.com/jingyaogong/minimind.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Compare commits

base: github:be0b8b4475f3406dd5eabc96169ab792e3774fa6
Branches Tags
github:master
github:website
github:v2
github:minimind-v1
...
compare: github:3064fdb90e67df9cce33612160a8b997090c2264
Branches Tags
github:master
github:website
github:v2
github:minimind-v1

8 Commits
be0b8b4475 ... 3064fdb90e

Author SHA1 Message Date
Jiyang 3064fdb90e
Merge 437eae89d6 into cc312c1cc6 2026-09-19 09:03:44 +08:00
jingyaogong cc312c1cc6 [fix] unify Ctrl+C handling 2026-09-18 16:33:09 +08:00
jingyaogong e2ee878519
Merge pull request #857 from wbbeyourself/fix-ctrl-c-exception 
fix: eval_llm.py Ctrl+C 退出异常修复
 2026-09-18 15:54:43 +08:00
jingyaogong b63a9565e3 [fix] restore top-1 router gradient #858 2026-09-18 15:28:50 +08:00
jingyaogong 1aae25911b
Merge pull request #859 from basil-k-aji-dev/fix/moe-router-gradient-topk1 
[fix] keep the MoE router trainable at num_experts_per_tok=1
 2026-09-18 15:03:36 +08:00
basil-k-aji-dev 1c876435ef [fix] keep the MoE router trainable at num_experts_per_tok=1 
Normalising a single top-k weight by its own sum gives exactly 1.0 for every
token, so the gate leaves the graph entirely: topk_idx is a discrete index and
the weight is now a constant, which leaves the LM loss no differentiable path
to the router. It then trains on aux_loss alone, whose optimum is a uniform
distribution, and the experts never specialise.

Measured on MOEFeedForward with router_aux_loss_coef=0, so only the LM path can
contribute a gradient:

  num_experts_per_tok=1, norm_topk_prob=True   7.02e-10   <- current default
  num_experts_per_tok=1, norm_topk_prob=False  1.02e-02
  num_experts_per_tok=2, norm_topk_prob=True   1.67e-02

Only normalise when there is more than one weight to normalise.

Note this changes the forward pass for existing num_experts_per_tok=1 MoE
checkpoints: the expert output was scaled by a constant 1.0 and is now scaled
by the gate probability.

Closes #858.
 2026-09-17 21:31:41 +05:30
wbbeyourself 9fa620e2e5 fix: eval_llm.py Ctrl+C 退出异常修复 2026-09-16 17:47:55 +08:00
Jiyang Liu 437eae89d6 [feat] add MPS support for Apple Silicon 
On macOS there's no CUDA, so these trainer scripts fall back to CPU. Detect MPS in the --device default and switch mixed precision to the device-agnostic torch.autocast / torch.amp.GradScaler so it works there.

CUDA and CPU paths are unchanged.
 2026-08-14 01:26:49 +08:00
10 changed files with 39 additions and 31 deletions
Show Stats Expand all files Collapse all files

5
eval_llm.py
View File

@ -91,4 +91,7 @@ def main():
print(f'\n[Speed]: {gen_tokens / (time.time() - st):.2f} tokens/s\n\n') if args.show_speed else print('\n\n')
if __name__ == "__main__":
main()
try:
main()
except (EOFError, KeyboardInterrupt):
print()

7
model/model_minimind.py
View File

@ -158,7 +158,12 @@ class MOEFeedForward(nn.Module):
x_flat = x.view(-1, hidden_dim)
scores = F.softmax(self.gate(x_flat), dim=-1)
topk_weight, topk_idx = torch.topk(scores, k=self.config.num_experts_per_tok, dim=-1, sorted=False)
if self.config.norm_topk_prob: topk_weight = topk_weight / (topk_weight.sum(dim=-1, keepdim=True) + 1e-20)
if self.config.norm_topk_prob:
if self.config.num_experts_per_tok > 1:
topk_weight = topk_weight / (topk_weight.sum(dim=-1, keepdim=True) + 1e-20)
else:
top1 = torch.topk(F.softmax(self.gate(x_flat.detach()), dim=-1), k=1, dim=-1, sorted=False)[0]
topk_weight = top1 - top1.detach() + 1.0 # k=1: 1.0 forward on purpose, gradient via straight-through
y = torch.zeros_like(x_flat)
for i, expert in enumerate(self.experts):
mask = (topk_idx == i)

6
trainer/train_agent.py
View File

@ -379,7 +379,7 @@ if __name__ == "__main__":
parser.add_argument("--epochs", type=int, default=1, help="训练轮数")
parser.add_argument("--batch_size", type=int, default=2, help="批次大小")
parser.add_argument("--learning_rate", type=float, default=3e-7, help="学习率")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else "cpu", help="训练设备")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else ("mps" if torch.backends.mps.is_available() else "cpu"), help="训练设备")
parser.add_argument("--dtype", type=str, default="bfloat16", help="数据类型 bfloat16/float16")
parser.add_argument("--num_workers", type=int, default=8, help="数据加载线程数")
parser.add_argument("--accumulation_steps", type=int, default=1, help="梯度累积步数")
@ -422,9 +422,9 @@ if __name__ == "__main__":
max_seq_len=args.max_seq_len + args.max_gen_len, use_moe=bool(args.use_moe))
ckp_data = lm_checkpoint(lm_config, weight=args.save_weight, save_dir='../checkpoints') if args.from_resume == 1 else None
device_type = "cuda" if "cuda" in args.device else "cpu"
device_type = "cuda" if "cuda" in args.device else args.device
dtype = torch.bfloat16 if args.dtype == "bfloat16" else torch.float16
autocast_ctx = nullcontext() if device_type == "cpu" else torch.cuda.amp.autocast(dtype=dtype)
autocast_ctx = nullcontext() if device_type == "cpu" else torch.autocast(device_type=device_type, dtype=dtype)
wandb = None
if args.use_wandb and is_main_process():

8
trainer/train_distillation.py
View File

@ -151,7 +151,7 @@ if __name__ == "__main__":
parser.add_argument("--epochs", type=int, default=6, help="训练轮数")
parser.add_argument("--batch_size", type=int, default=32, help="batch size")
parser.add_argument("--learning_rate", type=float, default=5e-6, help="初始学习率")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else "cpu", help="训练设备")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else ("mps" if torch.backends.mps.is_available() else "cpu"), help="训练设备")
parser.add_argument("--dtype", type=str, default="bfloat16", help="混合精度类型")
parser.add_argument("--num_workers", type=int, default=8, help="数据加载线程数")
parser.add_argument("--accumulation_steps", type=int, default=1, help="梯度累积步数")
@ -188,9 +188,9 @@ if __name__ == "__main__":
ckp_data = lm_checkpoint(lm_config_student, weight=args.save_weight, save_dir='../checkpoints') if args.from_resume==1 else None
# ========== 3. 设置混合精度 ==========
device_type = "cuda" if "cuda" in args.device else "cpu"
device_type = "cuda" if "cuda" in args.device else args.device
dtype = torch.bfloat16 if args.dtype == "bfloat16" else torch.float16
autocast_ctx = nullcontext() if device_type == "cpu" else torch.cuda.amp.autocast(dtype=dtype)
autocast_ctx = nullcontext() if device_type == "cpu" else torch.autocast(device_type=device_type, dtype=dtype)
# ========== 4. 配wandb ==========
wandb = None
@ -210,7 +210,7 @@ if __name__ == "__main__":
Logger(f'教师模型总参数量：{sum(p.numel() for p in teacher_model.parameters()) / 1e6:.3f} M')
train_ds = SFTDataset(args.data_path, tokenizer, max_length=args.max_seq_len)
train_sampler = DistributedSampler(train_ds) if dist.is_initialized() else None
scaler = torch.cuda.amp.GradScaler(enabled=(args.dtype == 'float16'))
scaler = torch.amp.GradScaler(device_type, enabled=(args.dtype == 'float16'))
optimizer = optim.AdamW(model.parameters(), lr=args.learning_rate)
# ========== 6. 从ckp恢复状态 ==========

8
trainer/train_dpo.py
View File

@ -135,7 +135,7 @@ if __name__ == "__main__":
parser.add_argument("--epochs", type=int, default=1, help="训练轮数")
parser.add_argument("--batch_size", type=int, default=4, help="batch size")
parser.add_argument("--learning_rate", type=float, default=4e-8, help="初始学习率（建议<=5e-8避免遗忘）")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else "cpu", help="训练设备")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else ("mps" if torch.backends.mps.is_available() else "cpu"), help="训练设备")
parser.add_argument("--dtype", type=str, default="bfloat16", help="混合精度类型")
parser.add_argument("--num_workers", type=int, default=8, help="数据加载线程数")
parser.add_argument("--accumulation_steps", type=int, default=1, help="梯度累积步数")
@ -166,9 +166,9 @@ if __name__ == "__main__":
ckp_data = lm_checkpoint(lm_config, weight=args.save_weight, save_dir='../checkpoints') if args.from_resume==1 else None
# ========== 3. 设置混合精度 ==========
device_type = "cuda" if "cuda" in args.device else "cpu"
device_type = "cuda" if "cuda" in args.device else args.device
dtype = torch.bfloat16 if args.dtype == "bfloat16" else torch.float16
autocast_ctx = nullcontext() if device_type == "cpu" else torch.cuda.amp.autocast(dtype=dtype)
autocast_ctx = nullcontext() if device_type == "cpu" else torch.autocast(device_type=device_type, dtype=dtype)
# ========== 4. 配wandb ==========
wandb = None
@ -190,7 +190,7 @@ if __name__ == "__main__":
train_ds = DPODataset(args.data_path, tokenizer, max_length=args.max_seq_len)
train_sampler = DistributedSampler(train_ds) if dist.is_initialized() else None
scaler = torch.cuda.amp.GradScaler(enabled=(args.dtype == 'float16'))
scaler = torch.amp.GradScaler(device_type, enabled=(args.dtype == 'float16'))
optimizer = optim.AdamW(model.parameters(), lr=args.learning_rate)
# ========== 6. 从ckp恢复状态 ==========

8
trainer/train_full_sft.py
View File

@ -88,7 +88,7 @@ if __name__ == "__main__":
parser.add_argument("--epochs", type=int, default=2, help="训练轮数")
parser.add_argument("--batch_size", type=int, default=16, help="batch size")
parser.add_argument("--learning_rate", type=float, default=1e-5, help="初始学习率")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else "cpu", help="训练设备")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else ("mps" if torch.backends.mps.is_available() else "cpu"), help="训练设备")
parser.add_argument("--dtype", type=str, default="bfloat16", help="混合精度类型")
parser.add_argument("--num_workers", type=int, default=8, help="数据加载线程数")
parser.add_argument("--accumulation_steps", type=int, default=1, help="梯度累积步数")
@ -119,9 +119,9 @@ if __name__ == "__main__":
ckp_data = lm_checkpoint(lm_config, weight=args.save_weight, save_dir='../checkpoints') if args.from_resume==1 else None
# ========== 3. 设置混合精度 ==========
device_type = "cuda" if "cuda" in args.device else "cpu"
device_type = "cuda" if "cuda" in args.device else args.device
dtype = torch.bfloat16 if args.dtype == "bfloat16" else torch.float16
autocast_ctx = nullcontext() if device_type == "cpu" else torch.cuda.amp.autocast(dtype=dtype)
autocast_ctx = nullcontext() if device_type == "cpu" else torch.autocast(device_type=device_type, dtype=dtype)
# ========== 4. 配wandb ==========
wandb = None
@ -136,7 +136,7 @@ if __name__ == "__main__":
model, tokenizer = init_model(lm_config, args.from_weight, device=args.device)
train_ds = SFTDataset(args.data_path, tokenizer, max_length=args.max_seq_len)
train_sampler = DistributedSampler(train_ds) if dist.is_initialized() else None
scaler = torch.cuda.amp.GradScaler(enabled=(args.dtype == 'float16'))
scaler = torch.amp.GradScaler(device_type, enabled=(args.dtype == 'float16'))
optimizer = optim.AdamW(model.parameters(), lr=args.learning_rate)
# ========== 6. 从ckp恢复状态 ==========

6
trainer/train_grpo.py
View File

@ -212,7 +212,7 @@ if __name__ == "__main__":
parser.add_argument("--epochs", type=int, default=1, help="训练轮数")
parser.add_argument("--batch_size", type=int, default=2, help="batch size")
parser.add_argument("--learning_rate", type=float, default=3e-7, help="初始学习率")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else "cpu", help="训练设备")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else ("mps" if torch.backends.mps.is_available() else "cpu"), help="训练设备")
parser.add_argument("--dtype", type=str, default="bfloat16", help="混合精度类型")
parser.add_argument("--num_workers", type=int, default=8, help="数据加载线程数")
parser.add_argument("--accumulation_steps", type=int, default=1, help="梯度累积步数")
@ -257,9 +257,9 @@ if __name__ == "__main__":
ckp_data = lm_checkpoint(lm_config, weight=args.save_weight, save_dir='../checkpoints') if args.from_resume==1 else None
# ========== 3. 设置混合精度 ==========
device_type = "cuda" if "cuda" in args.device else "cpu"
device_type = "cuda" if "cuda" in args.device else args.device
dtype = torch.bfloat16 if args.dtype == "bfloat16" else torch.float16
autocast_ctx = nullcontext() if device_type == "cpu" else torch.cuda.amp.autocast(dtype=dtype)
autocast_ctx = nullcontext() if device_type == "cpu" else torch.autocast(device_type=device_type, dtype=dtype)
# ========== 4. 配wandb ==========
wandb = None

8
trainer/train_lora.py
View File

@ -82,7 +82,7 @@ if __name__ == "__main__":
parser.add_argument("--epochs", type=int, default=10, help="训练轮数")
parser.add_argument("--batch_size", type=int, default=32, help="batch size")
parser.add_argument("--learning_rate", type=float, default=1e-4, help="初始学习率")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else "cpu", help="训练设备")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else ("mps" if torch.backends.mps.is_available() else "cpu"), help="训练设备")
parser.add_argument("--dtype", type=str, default="bfloat16", help="混合精度类型")
parser.add_argument("--num_workers", type=int, default=8, help="数据加载线程数")
parser.add_argument("--accumulation_steps", type=int, default=1, help="梯度累积步数")
@ -112,9 +112,9 @@ if __name__ == "__main__":
ckp_data = lm_checkpoint(lm_config, weight=args.lora_name, save_dir='../checkpoints') if args.from_resume==1 else None
# ========== 3. 设置混合精度 ==========
device_type = "cuda" if "cuda" in args.device else "cpu"
device_type = "cuda" if "cuda" in args.device else args.device
dtype = torch.bfloat16 if args.dtype == "bfloat16" else torch.float16
autocast_ctx = nullcontext() if device_type == "cpu" else torch.cuda.amp.autocast(dtype=dtype)
autocast_ctx = nullcontext() if device_type == "cpu" else torch.autocast(device_type=device_type, dtype=dtype)
# ========== 4. 配wandb ==========
wandb = None
@ -148,7 +148,7 @@ if __name__ == "__main__":
# ========== 6. 定义数据和优化器 ==========
train_ds = SFTDataset(args.data_path, tokenizer, max_length=args.max_seq_len)
train_sampler = DistributedSampler(train_ds) if dist.is_initialized() else None
scaler = torch.cuda.amp.GradScaler(enabled=(args.dtype == 'float16'))
scaler = torch.amp.GradScaler(device_type, enabled=(args.dtype == 'float16'))
optimizer = optim.AdamW(lora_params, lr=args.learning_rate)
# ========== 7. 从ckp恢复状态 ==========

6
trainer/train_ppo.py
View File

@ -315,7 +315,7 @@ if __name__ == "__main__":
parser.add_argument("--batch_size", type=int, default=2, help="batch size")
parser.add_argument("--learning_rate", type=float, default=3e-7, help="Actor学习率")
parser.add_argument("--critic_learning_rate", type=float, default=5e-7, help="Critic学习率")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else "cpu", help="训练设备")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else ("mps" if torch.backends.mps.is_available() else "cpu"), help="训练设备")
parser.add_argument("--dtype", type=str, default="bfloat16", help="混合精度类型")
parser.add_argument("--num_workers", type=int, default=8, help="数据加载线程数")
parser.add_argument("--accumulation_steps", type=int, default=1, help="梯度累积步数")
@ -364,9 +364,9 @@ if __name__ == "__main__":
ckp_data = lm_checkpoint(lm_config, weight=args.save_weight, save_dir='../checkpoints') if args.from_resume==1 else None
# ========== 3. 设置混合精度 ==========
device_type = "cuda" if "cuda" in args.device else "cpu"
device_type = "cuda" if "cuda" in args.device else args.device
dtype = torch.bfloat16 if args.dtype == "bfloat16" else torch.float16
autocast_ctx = nullcontext() if device_type == "cpu" else torch.cuda.amp.autocast(dtype=dtype)
autocast_ctx = nullcontext() if device_type == "cpu" else torch.autocast(device_type=device_type, dtype=dtype)
# ========== 4. 配wandb ==========
wandb = None

8
trainer/train_pretrain.py
View File

@ -87,7 +87,7 @@ if __name__ == "__main__":
parser.add_argument("--epochs", type=int, default=2, help="训练轮数")
parser.add_argument("--batch_size", type=int, default=32, help="batch size")
parser.add_argument("--learning_rate", type=float, default=5e-4, help="初始学习率")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else "cpu", help="训练设备")
parser.add_argument("--device", type=str, default="cuda:0" if torch.cuda.is_available() else ("mps" if torch.backends.mps.is_available() else "cpu"), help="训练设备")
parser.add_argument("--dtype", type=str, default="bfloat16", help="混合精度类型")
parser.add_argument("--num_workers", type=int, default=8, help="数据加载线程数")
parser.add_argument("--accumulation_steps", type=int, default=8, help="梯度累积步数")
@ -118,9 +118,9 @@ if __name__ == "__main__":
ckp_data = lm_checkpoint(lm_config, weight=args.save_weight, save_dir='../checkpoints') if args.from_resume==1 else None
# ========== 3. 设置混合精度 ==========
device_type = "cuda" if "cuda" in args.device else "cpu"
device_type = "cuda" if "cuda" in args.device else args.device
dtype = torch.bfloat16 if args.dtype == "bfloat16" else torch.float16
autocast_ctx = nullcontext() if device_type == "cpu" else torch.cuda.amp.autocast(dtype=dtype)
autocast_ctx = nullcontext() if device_type == "cpu" else torch.autocast(device_type=device_type, dtype=dtype)
# ========== 4. 配wandb ==========
wandb = None
@ -135,7 +135,7 @@ if __name__ == "__main__":
model, tokenizer = init_model(lm_config, args.from_weight, device=args.device)
train_ds = PretrainDataset(args.data_path, tokenizer, max_length=args.max_seq_len)
train_sampler = DistributedSampler(train_ds) if dist.is_initialized() else None
scaler = torch.cuda.amp.GradScaler(enabled=(args.dtype == 'float16'))
scaler = torch.amp.GradScaler(device_type, enabled=(args.dtype == 'float16'))
optimizer = optim.AdamW(model.parameters(), lr=args.learning_rate)
# ========== 6. 从ckp恢复状态 ==========