derek-thomas HF staff commited on
Commit
8039890
1 Parent(s): a5e29e6

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +8 -9
app.py CHANGED
@@ -123,15 +123,14 @@ with gr.Blocks(theme='NoCrypt/miku') as demo:
123
  """)
124
  with gr.Tab("Memory Calculation"):
125
  #with gr.TabItem("Memory Calculation"):
126
- with gr.Accordion("About Memory Calculation", open=False):
127
- gr.Markdown("""
128
- ## Memory Calculation
129
-
130
- Memory Calculation calculates the amount of device memory required to train or infer a model. See [Transformers Math 101](https://blog.eleuther.ai/transformer-math/) for more details on how memory overhead is calculated.
131
- Take this estimation with a grain of salt, because every implementation is different and these calculations were written to match the GPT-NeoX library as close as possible.
132
- Even for other training and inference libraries, however, we expect our script to give approximate memory estimations within acceptable error.
133
- (Please see [LLM finetuning memory requirements](https://blog.scottlogic.com/2023/11/24/llm-mem.html) for a treatment of how specific memory costs may vary framework-to-framework). Other good resources that we consulted are the [ZeRO Paper](https://arxiv.org/abs/1910.02054) and [Reducing Activation Recomputation in Large Transformer Models](https://arxiv.org/pdf/2205.05198.pdf).
134
- """)
135
  with gr.Accordion("How to use it?", open=False):
136
  gr.Markdown("""
137
  ## To Use
 
123
  """)
124
  with gr.Tab("Memory Calculation"):
125
  #with gr.TabItem("Memory Calculation"):
126
+ gr.Markdown("""
127
+ ## Memory Calculation
128
+
129
+ Memory Calculation calculates the amount of device memory required to train or infer a model. See [Transformers Math 101](https://blog.eleuther.ai/transformer-math/) for more details on how memory overhead is calculated.
130
+ Take this estimation with a grain of salt, because every implementation is different and these calculations were written to match the GPT-NeoX library as close as possible.
131
+ Even for other training and inference libraries, however, we expect our script to give approximate memory estimations within acceptable error.
132
+ (Please see [LLM finetuning memory requirements](https://blog.scottlogic.com/2023/11/24/llm-mem.html) for a treatment of how specific memory costs may vary framework-to-framework). Other good resources that we consulted are the [ZeRO Paper](https://arxiv.org/abs/1910.02054) and [Reducing Activation Recomputation in Large Transformer Models](https://arxiv.org/pdf/2205.05198.pdf).
133
+ """)
 
134
  with gr.Accordion("How to use it?", open=False):
135
  gr.Markdown("""
136
  ## To Use