Hugging Face's logo

NickyNicky
/
Llama-1B-base-GRPO-RAG-NEWS-SPANISH-v2

Text Generation
Transformers
Safetensors
Spanish
llama
finance
conversational
text-generation-inference
Model card Files Community
NickyNicky commited on
Commit
e0a4946
·
verified ·
1 Parent(s): 9492124

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +23 -1
README.md CHANGED
@@ -4,10 +4,32 @@ tags:
4
  - finance
5
  language:
6
  - es
 
 
 
 
7
  ---
8
 
9
 
10
- wow
 
11
 
12
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
13
  <!-- https://wandb.ai/multimodal_master/huggingface/runs/3fl5oqaw/workspace?nw=nwusertrainllms -->
 
4
  - finance
5
  language:
6
  - es
7
+ license: apache-2.0
8
+ datasets:
9
+ - NickyNicky/news_rag_spanish_14k_v1
10
+ pipeline_tag: text-generation
11
  ---
12
 
13
 
14
+ ## metrics (controlling sharp rises 'kl', 'loss' and 'grad_norm with 'bleu_reward_func').
15
+ ![image/png](https://cdn-uploads.huggingface.co/production/uploads/641b435ba5f876fe30c5ae0a/TBL0607_O8UTe4i721x9L.png)
16
 
17
 
18
+ ![image/png](https://cdn-uploads.huggingface.co/production/uploads/641b435ba5f876fe30c5ae0a/DvyKC-lcHFJrhnIlDYFCO.png)
19
+
20
+
21
+ ![image/png](https://cdn-uploads.huggingface.co/production/uploads/641b435ba5f876fe30c5ae0a/Mb5m_CCPx9mTMxzSiN4sw.png)
22
+
23
+
24
+ ![image/png](https://cdn-uploads.huggingface.co/production/uploads/641b435ba5f876fe30c5ae0a/lGACi-2AAN4ZRYLjsRD75.png)
25
+
26
+
27
+ note:
28
+ ```
29
+ I observe that in inferences and tests the model adapts more to the user's commands,
30
+ The responses are shorter depending on the 'prompt',
31
+ Responds better to Spanish and is more proficient in cryptocurrency news
32
+ knows more about interpreting what the user asks for in Spanish
33
+ ```
34
+
35
  <!-- https://wandb.ai/multimodal_master/huggingface/runs/3fl5oqaw/workspace?nw=nwusertrainllms -->