efederici commited on
Commit
3f94072
1 Parent(s): e6b1b73

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -4
README.md CHANGED
@@ -22,6 +22,9 @@ ipt-350m is:
22
  - **Capable of fast training and inference** (via [FlashAttention](https://arxiv.org/pdf/2205.14135.pdf) and [FasterTransformer](https://github.com/NVIDIA/FasterTransformer))
23
  - **Equipped with highly efficient open-source training code** via the [llm-foundry repository](https://github.com/mosaicml/llm-foundry)
24
 
 
 
 
25
  ## How to Use
26
 
27
  ```python
@@ -92,7 +95,4 @@ The model has been modified from a standard transformer in the following ways:
92
  The model was trained for ~13B tokens (with batch size 64 and sequence length 2048) on [OSCAR-2301](https://huggingface.co/datasets/oscar-corpus/OSCAR-2301).
93
  Each example was constructed from as many sequences from that dataset as were necessary to fill the 2048 sequence length.
94
 
95
- Vocabulary size is 50432, a multiple of 128 as suggested in [MEGATRON-LM](https://arxiv.org/abs/1909.08053), model flop utilization (MFU) increased by up to four percentage points.
96
-
97
- If you like this project, consider supporting me with a cup of coffee! 🤖✨🌞
98
- [![Buy me a coffee](https://badgen.net/badge/icon/Buy%20Me%20A%20Coffee?icon=buymeacoffee&label)](https://bmc.link/edoardofederici)
 
22
  - **Capable of fast training and inference** (via [FlashAttention](https://arxiv.org/pdf/2205.14135.pdf) and [FasterTransformer](https://github.com/NVIDIA/FasterTransformer))
23
  - **Equipped with highly efficient open-source training code** via the [llm-foundry repository](https://github.com/mosaicml/llm-foundry)
24
 
25
+ If you find this project useful, consider supporting its development:
26
+ [![Buy me a coffee](https://badgen.net/badge/icon/Buy%20Me%20A%20Coffee?icon=buymeacoffee&label)](https://bmc.link/edoardofederici)
27
+
28
  ## How to Use
29
 
30
  ```python
 
95
  The model was trained for ~13B tokens (with batch size 64 and sequence length 2048) on [OSCAR-2301](https://huggingface.co/datasets/oscar-corpus/OSCAR-2301).
96
  Each example was constructed from as many sequences from that dataset as were necessary to fill the 2048 sequence length.
97
 
98
+ Vocabulary size is 50432, a multiple of 128 as suggested in [MEGATRON-LM](https://arxiv.org/abs/1909.08053), model flop utilization (MFU) increased by up to four percentage points.