-
Mida õppisin uut tehniliste oskuste ja meeskonnatöö osas. Mis oli huvitav. Mille eest tahan kiita ennast ja mille eest kiidan teisi meeskonnaliikmeid. Mida võtan siit kaasa ja mida tahan juurde õppida…
-
Crawl data
1. Collect any data that suitable for our LLM
- Check quality of data // Should collect or not with the amount, available for collection, quality of text; paragraph, content, etc.
- C…
-
Do a literature review on the Multilingual Autoregressive model pretraining
Area to focus
1. Pretraining techniques that make multilingual training different from monolingual
2. Autoregressive models…
-
Please consider adding Jool package for AsusWRT-Merlin. There is Tayga package but it is not maintained anymore.
- Jool does NAT64/SIIT and is currently active and maintained.
- Website -- www.joo…
-
Explore pre-training configurations or other literatures
Example: Optimizer, dataset sampling, and vocab size of tokenizer
tokenizer type: SPM vs BPE
clean data thresold
learning rate (MUP)
!…
-
First, amazing theme, finally having something that works properly on my mobile device is so wonderful, thanks so much for developing this!
That said, and I hate to be that guy, but there are a cou…
-
Similar issue has been raised before, but this can not be worked around by running v6 before v4, as for instance with target MARK. The target is JOOL_SIIT
OS: Ubuntu 20.04 with kernel 5.4.0-104-gen…
-
**Description:**
Add MPT with Gradient Checkpointing and LoRa support into OpenThaiGPT pertaining code. We will use MPT with Lora for continue pertaining to task #179
**To Do:**
1. MPT Weight + MP…
-
### Description
Some software is IPv6-only. Docker should provide a SIIT-DC for running IPv6-only software on an IPv4-only host.
-
The doc says that for the tree:
```
root
|
+----A
| |
| +---B
| |
| +---C
|
+----D
|
+---E
|
+---F
```
The resulting order in which node…