Update README.md
Browse files
README.md
CHANGED
|
@@ -12,8 +12,8 @@ model-index:
|
|
| 12 |
- task:
|
| 13 |
type: text-generation
|
| 14 |
dataset:
|
| 15 |
-
|
| 16 |
-
|
| 17 |
metrics:
|
| 18 |
- name: pass@1
|
| 19 |
type: pass@1
|
|
@@ -22,8 +22,8 @@ model-index:
|
|
| 22 |
- task:
|
| 23 |
type: text-generation
|
| 24 |
dataset:
|
| 25 |
-
|
| 26 |
-
|
| 27 |
metrics:
|
| 28 |
- name: pass@1
|
| 29 |
type: pass@1
|
|
@@ -32,8 +32,8 @@ model-index:
|
|
| 32 |
- task:
|
| 33 |
type: text-generation
|
| 34 |
dataset:
|
| 35 |
-
|
| 36 |
-
|
| 37 |
metrics:
|
| 38 |
- name: pass@1
|
| 39 |
type: pass@1
|
|
@@ -42,8 +42,8 @@ model-index:
|
|
| 42 |
- task:
|
| 43 |
type: text-generation
|
| 44 |
dataset:
|
| 45 |
-
|
| 46 |
-
|
| 47 |
metrics:
|
| 48 |
- name: pass@1
|
| 49 |
type: pass@1
|
|
@@ -52,18 +52,18 @@ model-index:
|
|
| 52 |
- task:
|
| 53 |
type: text-generation
|
| 54 |
dataset:
|
| 55 |
-
|
| 56 |
-
|
| 57 |
metrics:
|
| 58 |
- name: pass@1
|
| 59 |
type: pass@1
|
| 60 |
-
value: 39
|
| 61 |
veriefied: false
|
| 62 |
- task:
|
| 63 |
type: text-generation
|
| 64 |
dataset:
|
| 65 |
-
|
| 66 |
-
|
| 67 |
metrics:
|
| 68 |
- name: pass@1
|
| 69 |
type: pass@1
|
|
@@ -72,8 +72,8 @@ model-index:
|
|
| 72 |
- task:
|
| 73 |
type: text-generation
|
| 74 |
dataset:
|
| 75 |
-
|
| 76 |
-
|
| 77 |
metrics:
|
| 78 |
- name: pass@1
|
| 79 |
type: pass@1
|
|
@@ -82,8 +82,8 @@ model-index:
|
|
| 82 |
- task:
|
| 83 |
type: text-generation
|
| 84 |
dataset:
|
| 85 |
-
|
| 86 |
-
|
| 87 |
metrics:
|
| 88 |
- name: pass@1
|
| 89 |
type: pass@1
|
|
@@ -92,8 +92,8 @@ model-index:
|
|
| 92 |
- task:
|
| 93 |
type: text-generation
|
| 94 |
dataset:
|
| 95 |
-
|
| 96 |
-
|
| 97 |
metrics:
|
| 98 |
- name: pass@1
|
| 99 |
type: pass@1
|
|
@@ -102,8 +102,8 @@ model-index:
|
|
| 102 |
- task:
|
| 103 |
type: text-generation
|
| 104 |
dataset:
|
| 105 |
-
|
| 106 |
-
|
| 107 |
metrics:
|
| 108 |
- name: pass@1
|
| 109 |
type: pass@1
|
|
@@ -112,8 +112,8 @@ model-index:
|
|
| 112 |
- task:
|
| 113 |
type: text-generation
|
| 114 |
dataset:
|
| 115 |
-
|
| 116 |
-
|
| 117 |
metrics:
|
| 118 |
- name: pass@1
|
| 119 |
type: pass@1
|
|
@@ -122,8 +122,8 @@ model-index:
|
|
| 122 |
- task:
|
| 123 |
type: text-generation
|
| 124 |
dataset:
|
| 125 |
-
|
| 126 |
-
|
| 127 |
metrics:
|
| 128 |
- name: pass@1
|
| 129 |
type: pass@1
|
|
@@ -132,8 +132,8 @@ model-index:
|
|
| 132 |
- task:
|
| 133 |
type: text-generation
|
| 134 |
dataset:
|
| 135 |
-
|
| 136 |
-
|
| 137 |
metrics:
|
| 138 |
- name: pass@1
|
| 139 |
type: pass@1
|
|
@@ -142,8 +142,8 @@ model-index:
|
|
| 142 |
- task:
|
| 143 |
type: text-generation
|
| 144 |
dataset:
|
| 145 |
-
|
| 146 |
-
|
| 147 |
metrics:
|
| 148 |
- name: pass@1
|
| 149 |
type: pass@1
|
|
@@ -152,8 +152,8 @@ model-index:
|
|
| 152 |
- task:
|
| 153 |
type: text-generation
|
| 154 |
dataset:
|
| 155 |
-
|
| 156 |
-
|
| 157 |
metrics:
|
| 158 |
- name: pass@1
|
| 159 |
type: pass@1
|
|
@@ -162,8 +162,8 @@ model-index:
|
|
| 162 |
- task:
|
| 163 |
type: text-generation
|
| 164 |
dataset:
|
| 165 |
-
|
| 166 |
-
|
| 167 |
metrics:
|
| 168 |
- name: pass@1
|
| 169 |
type: pass@1
|
|
@@ -172,33 +172,34 @@ model-index:
|
|
| 172 |
- task:
|
| 173 |
type: text-generation
|
| 174 |
dataset:
|
| 175 |
-
|
| 176 |
-
|
| 177 |
metrics:
|
| 178 |
- name: pass@1
|
| 179 |
type: pass@1
|
| 180 |
-
value: 23.
|
| 181 |
-
veriefied: false
|
| 182 |
- task:
|
| 183 |
type: text-generation
|
| 184 |
dataset:
|
| 185 |
-
|
| 186 |
-
|
| 187 |
metrics:
|
| 188 |
- name: pass@1
|
| 189 |
type: pass@1
|
| 190 |
value: 19.26
|
| 191 |
-
veriefied: false
|
| 192 |
- task:
|
| 193 |
type: text-generation
|
| 194 |
dataset:
|
| 195 |
-
|
| 196 |
-
|
| 197 |
metrics:
|
| 198 |
- name: pass@1
|
| 199 |
type: pass@1
|
| 200 |
value: 8.96
|
| 201 |
veriefied: false
|
|
|
|
| 202 |
---
|
| 203 |
|
| 204 |
<!--  -->
|
|
@@ -289,6 +290,11 @@ We train Granite 3.0 Language Models using IBM's super computing cluster, Blue V
|
|
| 289 |
**Ethical Considerations and Limitations:**
|
| 290 |
The use of Large Language Models involves risks and ethical considerations people must be aware of, including but not limited to: bias and fairness, misinformation, and autonomous decision-making. Granite-3.0-1B-A400M-Base model is not the exception in this regard. Even though this model is suited for multiple generative AI tasks, it has not undergone any safety alignment, there it may produce problematic outputs. Additionally, it remains uncertain whether smaller models might exhibit increased susceptibility to hallucination in generation scenarios by copying text verbatim from the training dataset due to their reduced sizes and memorization capacities. This aspect is currently an active area of research, and we anticipate more rigorous exploration, comprehension, and mitigations in this domain. Regarding ethics, a latent risk associated with all Large Language Models is their malicious utilization. We urge the community to use Granite-3.0-1B-A400M-Base model with ethical intentions and in a responsible way.
|
| 291 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 292 |
<!-- ## Citation
|
| 293 |
```
|
| 294 |
@misc{granite-models,
|
|
|
|
| 12 |
- task:
|
| 13 |
type: text-generation
|
| 14 |
dataset:
|
| 15 |
+
type: human-exams
|
| 16 |
+
name: MMLU
|
| 17 |
metrics:
|
| 18 |
- name: pass@1
|
| 19 |
type: pass@1
|
|
|
|
| 22 |
- task:
|
| 23 |
type: text-generation
|
| 24 |
dataset:
|
| 25 |
+
type: human-exams
|
| 26 |
+
name: MMLU-Pro
|
| 27 |
metrics:
|
| 28 |
- name: pass@1
|
| 29 |
type: pass@1
|
|
|
|
| 32 |
- task:
|
| 33 |
type: text-generation
|
| 34 |
dataset:
|
| 35 |
+
type: human-exams
|
| 36 |
+
name: AGI-Eval
|
| 37 |
metrics:
|
| 38 |
- name: pass@1
|
| 39 |
type: pass@1
|
|
|
|
| 42 |
- task:
|
| 43 |
type: text-generation
|
| 44 |
dataset:
|
| 45 |
+
type: commonsense
|
| 46 |
+
name: WinoGrande
|
| 47 |
metrics:
|
| 48 |
- name: pass@1
|
| 49 |
type: pass@1
|
|
|
|
| 52 |
- task:
|
| 53 |
type: text-generation
|
| 54 |
dataset:
|
| 55 |
+
type: commonsense
|
| 56 |
+
name: OBQA
|
| 57 |
metrics:
|
| 58 |
- name: pass@1
|
| 59 |
type: pass@1
|
| 60 |
+
value: 39
|
| 61 |
veriefied: false
|
| 62 |
- task:
|
| 63 |
type: text-generation
|
| 64 |
dataset:
|
| 65 |
+
type: commonsense
|
| 66 |
+
name: SIQA
|
| 67 |
metrics:
|
| 68 |
- name: pass@1
|
| 69 |
type: pass@1
|
|
|
|
| 72 |
- task:
|
| 73 |
type: text-generation
|
| 74 |
dataset:
|
| 75 |
+
type: commonsense
|
| 76 |
+
name: PIQA
|
| 77 |
metrics:
|
| 78 |
- name: pass@1
|
| 79 |
type: pass@1
|
|
|
|
| 82 |
- task:
|
| 83 |
type: text-generation
|
| 84 |
dataset:
|
| 85 |
+
type: commonsense
|
| 86 |
+
name: Hellaswag
|
| 87 |
metrics:
|
| 88 |
- name: pass@1
|
| 89 |
type: pass@1
|
|
|
|
| 92 |
- task:
|
| 93 |
type: text-generation
|
| 94 |
dataset:
|
| 95 |
+
type: commonsense
|
| 96 |
+
name: TruthfulQA
|
| 97 |
metrics:
|
| 98 |
- name: pass@1
|
| 99 |
type: pass@1
|
|
|
|
| 102 |
- task:
|
| 103 |
type: text-generation
|
| 104 |
dataset:
|
| 105 |
+
type: reading-comprehension
|
| 106 |
+
name: BoolQ
|
| 107 |
metrics:
|
| 108 |
- name: pass@1
|
| 109 |
type: pass@1
|
|
|
|
| 112 |
- task:
|
| 113 |
type: text-generation
|
| 114 |
dataset:
|
| 115 |
+
type: reading-comprehension
|
| 116 |
+
name: SQuAD 2.0
|
| 117 |
metrics:
|
| 118 |
- name: pass@1
|
| 119 |
type: pass@1
|
|
|
|
| 122 |
- task:
|
| 123 |
type: text-generation
|
| 124 |
dataset:
|
| 125 |
+
type: reasoning
|
| 126 |
+
name: ARC-C
|
| 127 |
metrics:
|
| 128 |
- name: pass@1
|
| 129 |
type: pass@1
|
|
|
|
| 132 |
- task:
|
| 133 |
type: text-generation
|
| 134 |
dataset:
|
| 135 |
+
type: reasoning
|
| 136 |
+
name: GPQA
|
| 137 |
metrics:
|
| 138 |
- name: pass@1
|
| 139 |
type: pass@1
|
|
|
|
| 142 |
- task:
|
| 143 |
type: text-generation
|
| 144 |
dataset:
|
| 145 |
+
type: reasoning
|
| 146 |
+
name: BBH
|
| 147 |
metrics:
|
| 148 |
- name: pass@1
|
| 149 |
type: pass@1
|
|
|
|
| 152 |
- task:
|
| 153 |
type: text-generation
|
| 154 |
dataset:
|
| 155 |
+
type: reasoning
|
| 156 |
+
name: MUSR
|
| 157 |
metrics:
|
| 158 |
- name: pass@1
|
| 159 |
type: pass@1
|
|
|
|
| 162 |
- task:
|
| 163 |
type: text-generation
|
| 164 |
dataset:
|
| 165 |
+
type: code
|
| 166 |
+
name: HumanEval
|
| 167 |
metrics:
|
| 168 |
- name: pass@1
|
| 169 |
type: pass@1
|
|
|
|
| 172 |
- task:
|
| 173 |
type: text-generation
|
| 174 |
dataset:
|
| 175 |
+
type: code
|
| 176 |
+
name: MBPP
|
| 177 |
metrics:
|
| 178 |
- name: pass@1
|
| 179 |
type: pass@1
|
| 180 |
+
value: 23.2
|
| 181 |
+
veriefied: false
|
| 182 |
- task:
|
| 183 |
type: text-generation
|
| 184 |
dataset:
|
| 185 |
+
type: math
|
| 186 |
+
name: GSM8K
|
| 187 |
metrics:
|
| 188 |
- name: pass@1
|
| 189 |
type: pass@1
|
| 190 |
value: 19.26
|
| 191 |
+
veriefied: false
|
| 192 |
- task:
|
| 193 |
type: text-generation
|
| 194 |
dataset:
|
| 195 |
+
type: math
|
| 196 |
+
name: MATH
|
| 197 |
metrics:
|
| 198 |
- name: pass@1
|
| 199 |
type: pass@1
|
| 200 |
value: 8.96
|
| 201 |
veriefied: false
|
| 202 |
+
new_version: ibm-granite/granite-3.1-1b-a400m-base
|
| 203 |
---
|
| 204 |
|
| 205 |
<!--  -->
|
|
|
|
| 290 |
**Ethical Considerations and Limitations:**
|
| 291 |
The use of Large Language Models involves risks and ethical considerations people must be aware of, including but not limited to: bias and fairness, misinformation, and autonomous decision-making. Granite-3.0-1B-A400M-Base model is not the exception in this regard. Even though this model is suited for multiple generative AI tasks, it has not undergone any safety alignment, there it may produce problematic outputs. Additionally, it remains uncertain whether smaller models might exhibit increased susceptibility to hallucination in generation scenarios by copying text verbatim from the training dataset due to their reduced sizes and memorization capacities. This aspect is currently an active area of research, and we anticipate more rigorous exploration, comprehension, and mitigations in this domain. Regarding ethics, a latent risk associated with all Large Language Models is their malicious utilization. We urge the community to use Granite-3.0-1B-A400M-Base model with ethical intentions and in a responsible way.
|
| 292 |
|
| 293 |
+
**Resources**
|
| 294 |
+
- ⭐️ Learn about the latest updates with Granite: https://www.ibm.com/granite
|
| 295 |
+
- 📄 Get started with tutorials, best practices, and prompt engineering advice: https://www.ibm.com/granite/docs/
|
| 296 |
+
- 💡 Learn about the latest Granite learning resources: https://ibm.biz/granite-learning-resources
|
| 297 |
+
|
| 298 |
<!-- ## Citation
|
| 299 |
```
|
| 300 |
@misc{granite-models,
|