Hugging Face's logo

HiTZ
/
latxa-7b-v1.2

Text Generation
Transformers
PyTorch
Basque
English
llama
Eval Results
text-generation-inference
Inference Endpoints
5 papers
Model card Files Community
OSainz commited on
Commit
5a798bd
•
1 Parent(s): 44608fd

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +11 -11
README.md CHANGED
@@ -21,7 +21,7 @@ model-index:
21
  metrics:
22
  - name: Accuracy (0-shot)
23
  type: Accuracy (0-shot)
24
- value: 65.72
25
  source:
26
  name: Paper
27
  url: https://arxiv.org/abs/2403.20266
@@ -33,7 +33,7 @@ model-index:
33
  metrics:
34
  - name: Accuracy (5-shot)
35
  type: Accuracy (5-shot)
36
- value: 36.89
37
  source:
38
  name: Paper
39
  url: https://arxiv.org/abs/2403.20266
@@ -45,7 +45,7 @@ model-index:
45
  metrics:
46
  - name: Average scores (5-shot)
47
  type: Average scores (5-shot)
48
- value: 51.78
49
  source:
50
  name: Paper
51
  url: https://arxiv.org/abs/2403.20266
@@ -57,7 +57,7 @@ model-index:
57
  metrics:
58
  - name: Accuracy (5-shot)
59
  type: Accuracy (5-shot)
60
- value: 32.44
61
  source:
62
  name: Paper
63
  url: https://arxiv.org/abs/2403.20266
@@ -69,7 +69,7 @@ model-index:
69
  metrics:
70
  - name: Accuracy (5-shot)
71
  type: Accuracy (5-shot)
72
- value: 30.40
73
  source:
74
  name: Paper
75
  url: https://arxiv.org/abs/2403.20266
@@ -81,7 +81,7 @@ model-index:
81
  metrics:
82
  - name: Accuracy (5-shot)
83
  type: Accuracy (5-shot)
84
- value: 44.37
85
  source:
86
  name: Paper
87
  url: https://arxiv.org/abs/2403.20266
@@ -93,7 +93,7 @@ model-index:
93
  metrics:
94
  - name: Accuracy (5-shot)
95
  type: Accuracy (5-shot)
96
- value: 34.20
97
  source:
98
  name: Paper
99
  url: https://arxiv.org/abs/2403.20266
@@ -251,18 +251,18 @@ In order to reproduce our results please follow the instructions in Latxa's [Git
251
  |
252
  | XGLM | 7B | 57.71 | 23.88 | 41.47 | 22.96 | 24.43 | 26.53 | 24.59 | 32.51 |
253
  | BLOOM | 7B | 57.18 | 27.00 | 40.17 | 25.34 | 28.41 | 27.17 | 25.07 | 33.86 |
254
- | Mistral | 7B | 51.09 | **38.89**| 39.22 | 25.01 | 29.26 | 34.58 | 32.15 | 35.94 |
255
  | Llama 2 | 7B | 50.43 | 26.22 | 38.20 | 24.09 | 27.27 | 29.50 | 28.84 | 32.51 |
256
- | **Latxa v1.1** | 7B | **65.72**| 36.89 | **51.78**| **32.44**| **30.40**| **44.37** | **34.20**| **42.26** |
257
  |
258
  | mGPT | 13B | 55.39 | 25.00 | 37.56 | 25.00 | 24.15 | 27.17 | 25.73 | 32.14 |
259
  | Llama 2 | 13B | 50.63 | 32.00 | 38.98 | 25.90 | 28.98 | 33.53 | 29.66 | 34.36 |
260
- | **Latxa v1.1** | 13B | **67.24**| 51.56 | 54.04 | **45.02**| 29.83 | **56.44** | **43.18**| 49.62 |
261
  |
262
  | Mixtral | 8x7B | 52.55 | 50.44 | 45.00 | 26.43 | 37.50 | 42.51 | 39.87 | 41.97 |
263
  | Yi | 34B | 52.22 | 54.56 | 43.90 | 27.30 | 34.66 | 42.57 | 39.68 | 42.05 |
264
  | Llama 2 | 70B | 51.62 | 33.56 | 42.55 | 24.16 | 27.84 | 38.43 | 33.08 | 35.47 |
265
- | **Latxa v1.1** | 70B | **69.76**| 64.89| **61.66**| **60.61**| **53.69**| **61.52** | **54.48**| **60.94** |
266
 
267
 
268
  # **Environmental Impact**
 
21
  metrics:
22
  - name: Accuracy (0-shot)
23
  type: Accuracy (0-shot)
24
+ value: 65.45
25
  source:
26
  name: Paper
27
  url: https://arxiv.org/abs/2403.20266
 
33
  metrics:
34
  - name: Accuracy (5-shot)
35
  type: Accuracy (5-shot)
36
+ value: 37.33
37
  source:
38
  name: Paper
39
  url: https://arxiv.org/abs/2403.20266
 
45
  metrics:
46
  - name: Average scores (5-shot)
47
  type: Average scores (5-shot)
48
+ value: 52.56
49
  source:
50
  name: Paper
51
  url: https://arxiv.org/abs/2403.20266
 
57
  metrics:
58
  - name: Accuracy (5-shot)
59
  type: Accuracy (5-shot)
60
+ value: 30.26
61
  source:
62
  name: Paper
63
  url: https://arxiv.org/abs/2403.20266
 
69
  metrics:
70
  - name: Accuracy (5-shot)
71
  type: Accuracy (5-shot)
72
+ value: 25.00
73
  source:
74
  name: Paper
75
  url: https://arxiv.org/abs/2403.20266
 
81
  metrics:
82
  - name: Accuracy (5-shot)
83
  type: Accuracy (5-shot)
84
+ value: 42.16
85
  source:
86
  name: Paper
87
  url: https://arxiv.org/abs/2403.20266
 
93
  metrics:
94
  - name: Accuracy (5-shot)
95
  type: Accuracy (5-shot)
96
+ value: 33.82
97
  source:
98
  name: Paper
99
  url: https://arxiv.org/abs/2403.20266
 
251
  |
252
  | XGLM | 7B | 57.71 | 23.88 | 41.47 | 22.96 | 24.43 | 26.53 | 24.59 | 32.51 |
253
  | BLOOM | 7B | 57.18 | 27.00 | 40.17 | 25.34 | 28.41 | 27.17 | 25.07 | 33.86 |
254
+ | Mistral | 7B | 51.09 | **38.89**| 39.22 | 25.01 | **29.26** | 34.58 | 32.15 | 35.94 |
255
  | Llama 2 | 7B | 50.43 | 26.22 | 38.20 | 24.09 | 27.27 | 29.50 | 28.84 | 32.51 |
256
+ | **Latxa v1.1** | 7B | **65.45**| 37.33 | **52.56**| **30.26**| 25.00| **42.16** | **33.82**| **40.94** |
257
  |
258
  | mGPT | 13B | 55.39 | 25.00 | 37.56 | 25.00 | 24.15 | 27.17 | 25.73 | 32.14 |
259
  | Llama 2 | 13B | 50.63 | 32.00 | 38.98 | 25.90 | 28.98 | 33.53 | 29.66 | 34.36 |
260
+ | **Latxa v1.1** | 13B | **66.51**| **53.89** | **53.36** | **44.11**| **32.67** | **56.38** | **43.66**| **50.08** |
261
  |
262
  | Mixtral | 8x7B | 52.55 | 50.44 | 45.00 | 26.43 | 37.50 | 42.51 | 39.87 | 41.97 |
263
  | Yi | 34B | 52.22 | 54.56 | 43.90 | 27.30 | 34.66 | 42.57 | 39.68 | 42.05 |
264
  | Llama 2 | 70B | 51.62 | 33.56 | 42.55 | 24.16 | 27.84 | 38.43 | 33.08 | 35.47 |
265
+ | **Latxa v1.1** | 70B | **70.55**| **71.67** | **59.74**| **60.65**| **50.57**| **62.45** | **51.90**| **61.08** |
266
 
267
 
268
  # **Environmental Impact**