DogeWatch

Results 8 comments of DogeWatch

v1.1版本deepspeed微调报错RuntimeError: output tensor must have the same type as input tensor

I have the same problem, https://github.com/microsoft/DeepSpeed/issues/3654

Add line break support for paragraphs

any updates on this issue?

[BUG]RuntimeError: output tensor must have the same type as input tensor

> @DogeWatch the format for enabling bf16 in your config is incorrect it should be `"bf16": { "enabled": true },` @jomayeri sorry, it's my mistake, but with "bf16" it still...

[BUG]NotImplementedError: You must implement either the backward or vjp method for your custom autograd.Function to use it with backward mode AD.

@PanQiWei yes, it works

读不舍手导入后提示书源解析失败

又提示解析失败了

GRPO训练报错：Fatal Python error: none_dealloc: deallocating None: bug likely caused by a refcount error in a C extension

训练了6k步后遇到了类似的报错 ``` Fatal Python error: Fatal Python error: none_dealloc: none_dealloc: deallocating None Fatal Python error: deallocating None Python runtime state: initializednone_dealloc: Python runtime state: initialized deallocating None Thread 0x00007f3b6f3f6700Python runtime...

GRPO训练报错：Fatal Python error: none_dealloc: deallocating None: bug likely caused by a refcount error in a C extension

> The code for GRPOTrainer has been refactored. Please try again using the main branch. v3.4.1 这个版本修复了吗？

Can the entire conversation round from the agent be outputted directly to the answer node in Chatflow?

> Are there any specific scenarios that need to use this feature? users want to know all the details of the agent, including the tools input and output. This would...