yujunjun/LLMs-from-scratch
1
0
Fork 0
mirror of https://github.com/rasbt/LLMs-from-scratch.git synced 2025-08-30 11:31:08 +00:00
Code Issues Packages Projects Releases Wiki Activity

add usage

Browse Source
This commit is contained in:
rasbt 2024-04-24 07:27:04 -05:00
parent fb54b064c9
commit d311bae25a
1 changed files with 16 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
ch06/02_bonus_additional-experiments/README.md
View File

@ -7,6 +7,8 @@ For example,
- comparing rows 1 and 3 answers the question: "What is the performance difference when we train only the last layer instead of the last block?";
- and so forth.
 
| | Model | Weights | Trainable token | Trainable layers | Context length | CPU/GPU | Training time | Training acc | Validation acc | Test acc |
|---|--------------------|------------|-----------------|------------------|-------------------------|---------|---------------|--------------|----------------|----------|
| 1 | gpt2-small (124M) | pretrained | last | last_block | longest train ex. (120) | V100 | 0.39 min | 96.63% | 97.99% | 94.33% |
@ -17,3 +19,16 @@ For example,
| 6 | gpt2-large (774M) | pretrained | last | last_block | longest train ex. (120) | V100 | 1.91 min | 99.52% | 98.66% | 96.67% |
| 7 | gpt2-small (124M) | random | last | all | longest train ex. (120) | V100 | 0.93 min | 100% | 97.32% | 93.00% |
| 8 | gpt2-small (124M) | pretrained | last | last_block | context length (1024) | V100 | 3.24 min | 83.08% | 87.92% | 78.33% |
 
### Usage:
- Row 1: `python additional-experiments.py`
- Row 2: `python additional-experiments.py --trainable_token first`
- Row 3: `python additional-experiments.py --trainable_layers last_layer`
- Row 4: `python additional-experiments.py --trainable_layers all`
- Row 5: `python additional-experiments.py --model_size gpt2-medium (355M)`
- Row 6: `python additional-experiments.py --model_size gpt2-large (774M)`
- Row 7: `python additional-experiments.py --weights random --trainable_layers all`
- Row 8: `python additional-experiments.py --context_length "model_context_length"`