File size: 10,933 Bytes
ee7140d d3527c7 603d4fa ee7140d 084b2ec 603d4fa 92e3107 63fb4d6 92e3107 48ecb9b 084b2ec |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 |
---
language:
- multilingual
- en
- ru
- zh
- de
- es
- fr
- ja
- it
- pt
- el
- ko
- fi
- id
- tr
- ar
- vi
- th
- bg
- ca
- hi
- et
- bn
- ta
- ur
- sw
- te
- eu
- my
- ht
- qu
license: mit
thumbnail: https://huggingface.co/front/thumbnails/facebook.png
inference: false
model-index:
- name: xglm-564M
results:
- task:
type: text-generation
name: Text Generation
dataset:
name: AI2 Reasoning Challenge (25-Shot)
type: ai2_arc
config: ARC-Challenge
split: test
args:
num_few_shot: 25
metrics:
- type: acc_norm
value: 24.57
name: normalized accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=facebook/xglm-564M
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: HellaSwag (10-Shot)
type: hellaswag
split: validation
args:
num_few_shot: 10
metrics:
- type: acc_norm
value: 34.64
name: normalized accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=facebook/xglm-564M
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: MMLU (5-Shot)
type: cais/mmlu
config: all
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 25.18
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=facebook/xglm-564M
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: TruthfulQA (0-shot)
type: truthful_qa
config: multiple_choice
split: validation
args:
num_few_shot: 0
metrics:
- type: mc2
value: 40.43
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=facebook/xglm-564M
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: Winogrande (5-shot)
type: winogrande
config: winogrande_xl
split: validation
args:
num_few_shot: 5
metrics:
- type: acc
value: 52.25
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=facebook/xglm-564M
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: GSM8k (5-shot)
type: gsm8k
config: main
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 0.23
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=facebook/xglm-564M
name: Open LLM Leaderboard
---
# XGLM-564M
XGLM-564M is a multilingual autoregressive language model (with 564 million parameters) trained on a balanced corpus of a diverse set of 30 languages totaling 500 billion sub-tokens. It was introduced in the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin\*, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li\* (\*Equal Contribution). The original implementation was released in [this repository](https://github.com/pytorch/fairseq/tree/main/examples/xglm).
## Training Data Statistics
The training data statistics of XGLM-564M is shown in the table below.
| ISO-639-1| family | name | # tokens | ratio | ratio w/ lowRes upsampling |
|:--------|:-----------------|:------------------------|-------------:|------------:|-------------:|
| en | Indo-European | English | 803526736124 | 0.489906 | 0.3259 |
| ru | Indo-European | Russian | 147791898098 | 0.0901079 | 0.0602 |
| zh | Sino-Tibetan | Chinese | 132770494630 | 0.0809494 | 0.0483 |
| de | Indo-European | German | 89223707856 | 0.0543992 | 0.0363 |
| es | Indo-European | Spanish | 87303083105 | 0.0532282 | 0.0353 |
| fr | Indo-European | French | 77419639775 | 0.0472023 | 0.0313 |
| ja | Japonic | Japanese | 66054364513 | 0.040273 | 0.0269 |
| it | Indo-European | Italian | 41930465338 | 0.0255648 | 0.0171 |
| pt | Indo-European | Portuguese | 36586032444 | 0.0223063 | 0.0297 |
| el | Indo-European | Greek (modern) | 28762166159 | 0.0175361 | 0.0233 |
| ko | Koreanic | Korean | 20002244535 | 0.0121953 | 0.0811 |
| fi | Uralic | Finnish | 16804309722 | 0.0102455 | 0.0681 |
| id | Austronesian | Indonesian | 15423541953 | 0.00940365 | 0.0125 |
| tr | Turkic | Turkish | 12413166065 | 0.00756824 | 0.0101 |
| ar | Afro-Asiatic | Arabic | 12248607345 | 0.00746791 | 0.0099 |
| vi | Austroasiatic | Vietnamese | 11199121869 | 0.00682804 | 0.0091 |
| th | Tai–Kadai | Thai | 10842172807 | 0.00661041 | 0.044 |
| bg | Indo-European | Bulgarian | 9703797869 | 0.00591635 | 0.0393 |
| ca | Indo-European | Catalan | 7075834775 | 0.0043141 | 0.0287 |
| hi | Indo-European | Hindi | 3448390110 | 0.00210246 | 0.014 |
| et | Uralic | Estonian | 3286873851 | 0.00200399 | 0.0133 |
| bn | Indo-European | Bengali, Bangla | 1627447450 | 0.000992245 | 0.0066 |
| ta | Dravidian | Tamil | 1476973397 | 0.000900502 | 0.006 |
| ur | Indo-European | Urdu | 1351891969 | 0.000824241 | 0.0055 |
| sw | Niger–Congo | Swahili | 907516139 | 0.000553307 | 0.0037 |
| te | Dravidian | Telugu | 689316485 | 0.000420272 | 0.0028 |
| eu | Language isolate | Basque | 105304423 | 6.42035e-05 | 0.0043 |
| my | Sino-Tibetan | Burmese | 101358331 | 6.17976e-05 | 0.003 |
| ht | Creole | Haitian, Haitian Creole | 86584697 | 5.27902e-05 | 0.0035 |
| qu | Quechuan | Quechua | 3236108 | 1.97304e-06 | 0.0001 |
## Model card
For intended usage of the model, please refer to the [model card](https://github.com/pytorch/fairseq/blob/main/examples/xglm/model_card.md) released by the XGLM-564M development team.
## Example (COPA)
The following snippet shows how to evaluate our models (GPT-3 style, zero-shot) on the Choice of Plausible Alternatives (COPA) task, using examples in English, Chinese and Hindi.
```python
import torch
import torch.nn.functional as F
from transformers import XGLMTokenizer, XGLMForCausalLM
tokenizer = XGLMTokenizer.from_pretrained("facebook/xglm-564M")
model = XGLMForCausalLM.from_pretrained("facebook/xglm-564M")
data_samples = {
'en': [
{
"premise": "I wanted to conserve energy.",
"choice1": "I swept the floor in the unoccupied room.",
"choice2": "I shut off the light in the unoccupied room.",
"question": "effect",
"label": "1"
},
{
"premise": "The flame on the candle went out.",
"choice1": "I blew on the wick.",
"choice2": "I put a match to the wick.",
"question": "cause",
"label": "0"
}
],
'zh': [
{
"premise": "我想节约能源。",
"choice1": "我在空着的房间里扫了地板。",
"choice2": "我把空房间里的灯关了。",
"question": "effect",
"label": "1"
},
{
"premise": "蜡烛上的火焰熄灭了。",
"choice1": "我吹灭了灯芯。",
"choice2": "我把一根火柴放在灯芯上。",
"question": "cause",
"label": "0"
}
],
'hi': [
{
"premise": "M te vle konsève enèji.",
"choice1": "Mwen te fin baleye chanm lib la.",
"choice2": "Mwen te femen limyè nan chanm lib la.",
"question": "effect",
"label": "1"
},
{
"premise": "Flam bouji a te etenn.",
"choice1": "Mwen te soufle bouji a.",
"choice2": "Mwen te limen mèch bouji a.",
"question": "cause",
"label": "0"
}
]
}
def get_logprobs(prompt):
inputs = tokenizer(prompt, return_tensors="pt")
input_ids, output_ids = inputs["input_ids"], inputs["input_ids"][:, 1:]
outputs = model(**inputs, labels=input_ids)
logits = outputs.logits
logprobs = torch.gather(F.log_softmax(logits, dim=2), 2, output_ids.unsqueeze(2))
return logprobs
# Zero-shot evaluation for the Choice of Plausible Alternatives (COPA) task.
# A return value of 0 indicates that the first alternative is more plausible,
# while 1 indicates that the second alternative is more plausible.
def COPA_eval(prompt, alternative1, alternative2):
lprob1 = get_logprobs(prompt + "\n" + alternative1).sum()
lprob2 = get_logprobs(prompt + "\n" + alternative2).sum()
return 0 if lprob1 > lprob2 else 1
for lang in data_samples_long:
for idx, example in enumerate(data_samples_long[lang]):
predict = COPA_eval(example["premise"], example["choice1"], example["choice2"])
print(f'{lang}-{idx}', predict, example['label'])
# en-0 1 1
# en-1 0 0
# zh-0 1 1
# zh-1 0 0
# hi-0 1 1
# hi-1 0 0
```
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_facebook__xglm-564M)
| Metric |Value|
|---------------------------------|----:|
|Avg. |29.55|
|AI2 Reasoning Challenge (25-Shot)|24.57|
|HellaSwag (10-Shot) |34.64|
|MMLU (5-Shot) |25.18|
|TruthfulQA (0-shot) |40.43|
|Winogrande (5-shot) |52.25|
|GSM8k (5-shot) | 0.23|
|