卷积计算的思考
下面这个卷积网络,的算力需求大概是多少 Mops , 权重是多少K 个
注意 stride , pooling 以及 深度可分离卷积 的参数
请直接输出一个表格, 输入是 16000 , 分析每一层的输出尺寸
根据你提供的配置参数,这是一个典型的 1D 卷积神经网络(常用于音频原始波形处理)。
conv_dim: Tuple[int, ...] = (8, 16, 32, 64, 128, 256, 16) # Progressive increase to 512
conv_kernel: Tuple[int, ...] = (10, 3, 3, 3, 3, 2, 2) # First layer larger kernel
conv_stride: Tuple[int, ...] = (5, 2, 2, 2, 2, 2, 2) # First layer larger stride
conv_bias: Tuple[bool, ...] = (False, False, False, False, False, False, False)
conv_padding: Tuple[int, ...] = (0, 0, 0, 0, 0, 0, 0)
# Disable depthwise for stability
conv_use_depthwise: Tuple[bool, ...] = (False, False, False, False, False, False, False)
# Reduce pooling - only 2 layers
conv_use_pool: Tuple[bool, ...] = (False, False, False, False, False, False, False)
# Disable residual for now (can cause gradient issues)
conv_use_residual: Tuple[bool, ...] = (False, False, False, False, False, False, False)
conv_pool_type: Tuple[str, ...] = ("avg", "avg", "avg", "avg", "avg")
conv_pool_kernel: Tuple[int, ...] = (2, 2, 2, 2, 2)
conv_pool_stride: Tuple[int, ...] = (2, 2, 2, 2, 2)
conv_pool_padding: Tuple[int, ...] = (0, 0, 0, 0, 0)
不思考,不联网
- 豆包 20260228:106.45K 136.01Mops
- Hunyuan 20260228:2440 K 3722 MOPs
- DeepseekV3.2:98,448 ~1.59 MOPs
- GPT-5.2:106,448 16,324,528MACs 32.649 Mops
- GLM5:106.42 K 16,717.69 M
- Qwen3.5-Plus:139.2 K (139,216 个参数) 约 19.7 M (19,712,000 次乘加运算)
- kimi k2.5:106.688 K (约 107K) 10.99 Mops (约 11 Mops)
- Claude-Haiku-4.5:~106.4K个参数 ~620.5 MOps(在32位浮点下)
- Claude-Opus-4.5:106.45 K 16.33 MOps
思考
- Gemini-3-Flash:106.45 K 32.66 Mops
- DeepseekV3.2:~105.4 K ~58.1 Mops
- 豆包 20260228:106.45 K(千个) 16.33 Mops
- Minimax-M2.5:106448(约104K)16328752(约16.33M)
No comments to display
No comments to display