Mars

https://github.com/NATSpeech/NATSpeech/blob/2e7084e4c76ee165de8d6ff7dacf6011514fbe7c/modules/tts/portaspeech/fvae.py#L128 请问一下，为什么这里直接从正态分布中采样，而不是通过vae的encoder预测除μ，σ。然后z*sigma+μ呢？

https://github.com/NATSpeech/NATSpeech/blob/main/modules/tts/portaspeech/fvae.py#L120 我的理解是： 这个loss是计算VAE生成的zq在VAE生成的分布q_dist中的概率，和通过prior_flow转换zq生成的zp在正态分布中的概率，之间的KL距离更小。如果KL距离很小，说明VPFlow转换的分布和VAE的分布相近。 但是怎么证明VPFlow转换的zp是正态分布呢？是不是还应该保证logpx和logqx足够高呢？

请问怎么修改配置问下，可以解决这个问题？ `Traceback (most recent call last): File "tasks/run.py", line 19, in run_task() File "tasks/run.py", line 14, in run_task task_cls.start() File "/cpfs01/shared/public/msm/workspace/NATSpeech/utils/commons/base_task.py", line 227, in start trainer.fit(cls) File "/cpfs01/shared/public/msm/workspace/NATSpeech/utils/commons/trainer.py", line 122,...

pretrain loss

4

Excuse me, what value does my pre-training loss reach, can I start fintune tts? i found my finued tts model can generate a mel-spectrom but diffrent to ori mel-spectrom very...

decoder输出长度是有限制吗？

5

parser.add_argument('--base_model', default="llama-2-7b-chat-hf/", type=str) parser.add_argument('--lora_weights', default="tloen/alpaca-lora-7b", type=str, help="If None, perform inference on the base model") parser.add_argument('--load_8bit', default="True", type=bool, help='only use CPU for inference') `You are using the default legacy behaviour of...

stream_output 是不是没用到呀

1

https://github.com/git-cloner/llama2-lora-fine-tuning/blob/98344720925c832142fb4f59b587231fd5496965/generate.py#L91

finetune error

2

`Loading checkpoint shards: 100%|█████████████████████████████████████████████████████████████████| 2/2 [00:04

2 node, tp1pp1ep16, mixtral ` source /home/aigc/miniforge3/bin/activate mamba GPUS_PER_NODE=8 export CUDA_DEVICE_MAX_CONNECTIONS=1 export NVSHMEM_DIR=/opt/nvshmem_src/ # Use for DeepEP installation export LD_LIBRARY_PATH="${NVSHMEM_DIR}/lib:$LD_LIBRARY_PATH" export PATH="${NVSHMEM_DIR}/bin:$PATH" export NCCL_DEBUG=INFO export NCCL_IB_DISABLE=0 export NCCL_SOCKET_IFNAME=eth0 export NCCL_IB_HCA=mlx5_0:1,mlx5_1:1,mlx5_2:1,mlx5_3:1,mlx5_6:1,mlx5_7:1...

error when testing test_internode.sh deep_ep.cpp:83 'an illegal memory access was encountered'

8

[testing] Running with BF16, without top-k (async=False, previous=False) ... passed [testing] Running with BF16, with top-k (async=False, previous=False) ... passed [testing] Running with BF16, without top-k (async=False, previous=False) ... passed...