Text Generation
Transformers
Safetensors
Czech
llama
text-generation-inference
Inference Endpoints
mfajcik commited on
Commit
8c8d368
·
verified ·
1 Parent(s): b3cba5a

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -0
README.md CHANGED
@@ -28,6 +28,7 @@ Below we
28
 
29
  ## Distance in y Between Fine-Tuning and Training from Scratch
30
  <img src="figures/tllama_test_distance.png" width="900"/>
 
31
 
32
  ## Training parameters
33
  Not mentioned parameters are the same as for [TinyLLama-2.5T](https://huggingface.co/TinyLlama/TinyLlama-1.1B-intermediate-step-1195k-token-2.5T).
 
28
 
29
  ## Distance in y Between Fine-Tuning and Training from Scratch
30
  <img src="figures/tllama_test_distance.png" width="900"/>
31
+ The distance |x1-x2| with same function value f1(x1)=f2(x2) grows with more steps. On convergence, it starts to rapidly increase (perhaps exponentially).
32
 
33
  ## Training parameters
34
  Not mentioned parameters are the same as for [TinyLLama-2.5T](https://huggingface.co/TinyLlama/TinyLlama-1.1B-intermediate-step-1195k-token-2.5T).