-
# 中世纪文学 Le moyen âge
- temp：12-15, s'étend du XIIe siècle à la fin du XVe
- arrière-plan：
- la monarchie s'impose
- 1163，notre-dame de paris
- 1257，la sorbonne, les premières univisités
…
aJiea updated
4 years ago
-
Typical use case: watching users' activities on the GitHub
ではUbuntuの場合の説明しか記述されていないため、
macなどでやろうとした場合にFluentdの導入でつまづいてしまう
sesta updated
9 years ago
-
### Describe the bug
I installed text generation webui and downloaded the model(TheBloke_Yarn-Mistral-7B-128k-AWQ) and I can't run it. I chose Transofmer as Model loader. I tried installing autoawq b…
-
I am having trouble in getting the benchmarks running with XLA backend. Used latest pytorch master/release2.0 branch.
Here is what I did
#!/bin/bash
set -x
# Setup the output directory
backend="t…
-
Hi!
Seems that the Ruby update to 2.6.x caused a problem when you try to run `systemctl reload google-fluentd`. We get this error with the latest `google-fluentd` releases:
```
2020-01-21 00:00…
-
Currently several schemes of sparse attention (e.g. block-sparse, sliding window) relies on the handcrafted kernels, and it takes plenty of effort to implement new schemes (for research or other purpo…
-
## 一开始照搬模型设置训练了一个大型数据集，始终无法收敛到理想区间，又拿斗破来修改模型参数玩了个把星期，各种调参。
### 总结如下：
#### 1. 模型的收敛取决于词嵌入的维度，维度越大收敛越快越好。（有没有上限就懒得去测试了，电费要紧。）
#### 2.head与隐藏层数可以适当裁剪，隐藏层可以设置高一些，multi-head感觉超过5层之后似乎对于生成的结果影响并不大。
…
-
## 🐛 Bug
I try to using amp in Pytorch core with torch.nn.DataParallel for multi-gpu training. I wrap forward pass in model in autocast, but get error
## To Reproduce
Steps to reproduce the…
-
### 🐛 Describe the bug
Repro:
```
import torch
import logging
import sys
import torch._dynamo
# torch._logging.set_logs(dynamo=logging.DEBUG, bytecode=True)
torch._dynamo.config.print_graph_…
-
### Description
Trax is a library for deep learning that focuses on sequence models and reinforcement learning. It combines performance with code clarity and maintained documentation and tests.
...
…