haoranxu commited on
Commit
f4da71e
1 Parent(s): cd195b3

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +15 -2
README.md CHANGED
@@ -3,6 +3,9 @@ license: mit
3
  ---
4
  **ALMA** (**A**dvanced **L**anguage **M**odel-based tr**A**nslator) is an LLM-based translation model, which adopts a new translation model paradigm: it begins with fine-tuning on monolingual data and is further optimized using high-quality parallel data. This two-step fine-tuning process ensures strong translation performance.
5
  Please find more details in our [paper](https://arxiv.org/abs/2309.11674).
 
 
 
6
  ```
7
  @misc{xu2023paradigm,
8
  title={A Paradigm Shift in Machine Translation: Boosting Translation Performance of Large Language Models},
@@ -13,21 +16,31 @@ Please find more details in our [paper](https://arxiv.org/abs/2309.11674).
13
  primaryClass={cs.CL}
14
  }
15
  ```
16
- We release four translation models presented in the paper:
17
  - **ALMA-7B**: Full-weight Fine-tune LLaMA-2-7B on 20B monolingual tokens and then **Full-weight** fine-tune on human-written parallel data
18
  - **ALMA-7B-LoRA**: Full-weight Fine-tune LLaMA-2-7B on 20B monolingual tokens and then **LoRA** fine-tune on human-written parallel data
 
19
  - **ALMA-13B**: Full-weight Fine-tune LLaMA-2-7B on 12B monolingual tokens and then **Full-weight** fine-tune on human-written parallel data
20
  - **ALMA-13B-LoRA** (Our best system): Full-weight Fine-tune LLaMA-2-7B on 12B monolingual tokens and then **LoRA** fine-tune on human-written parallel data
 
21
 
22
  Model checkpoints are released at huggingface:
23
  | Models | Base Model Link | LoRA Link |
24
  |:-------------:|:---------------:|:---------:|
25
  | ALMA-7B | [haoranxu/ALMA-7B](https://huggingface.co/haoranxu/ALMA-7B) | - |
26
  | ALMA-7B-LoRA | [haoranxu/ALMA-7B-Pretrain](https://huggingface.co/haoranxu/ALMA-7B-Pretrain) | [haoranxu/ALMA-7B-Pretrain-LoRA](https://huggingface.co/haoranxu/ALMA-7B-Pretrain-LoRA) |
 
27
  | ALMA-13B | [haoranxu/ALMA-13B](https://huggingface.co/haoranxu/ALMA-13B) | - |
28
  | ALMA-13B-LoRA | [haoranxu/ALMA-13B-Pretrain](https://huggingface.co/haoranxu/ALMA-13B-Pretrain) | [haoranxu/ALMA-13B-Pretrain-LoRA](https://huggingface.co/haoranxu/ALMA-13B-Pretrain-LoRA) |
 
29
 
30
- **Note that `ALMA-7B-Pretrain` and `ALMA-13B-Pretrain` are NOT translation models. They only experience stage 1 monolingual fine-tuning (20B tokens for the 7B model and 12B tokens for the 13B model), and should be utilized in conjunction with their LoRA models for translation purposes.**
 
 
 
 
 
 
31
 
32
  A quick start to use our best system (ALMA-13B-LoRA) for translation. An example of translating "我爱机器翻译。" into English:
33
  ```
 
3
  ---
4
  **ALMA** (**A**dvanced **L**anguage **M**odel-based tr**A**nslator) is an LLM-based translation model, which adopts a new translation model paradigm: it begins with fine-tuning on monolingual data and is further optimized using high-quality parallel data. This two-step fine-tuning process ensures strong translation performance.
5
  Please find more details in our [paper](https://arxiv.org/abs/2309.11674).
6
+
7
+ **[ALMA-R](https://arxiv.org/abs/2401.08417) (NEW!) is released now!** ALMA-R builds upon ALMA models, with further LoRA fine-tuning with our proposed **Contrastive Preference Optimization (CPO)** as opposed to the Supervised Fine-tuning used in ALMA. CPO fine-tuning requires our [triplet preference data](https://huggingface.co/datasets/haoranxu/ALMA-R-Preference) for preference learning. ALMA-R now can matches or even exceeds GPT-4 or WMT winners!
8
+
9
  ```
10
  @misc{xu2023paradigm,
11
  title={A Paradigm Shift in Machine Translation: Boosting Translation Performance of Large Language Models},
 
16
  primaryClass={cs.CL}
17
  }
18
  ```
19
+ We release six translation models presented in the paper:
20
  - **ALMA-7B**: Full-weight Fine-tune LLaMA-2-7B on 20B monolingual tokens and then **Full-weight** fine-tune on human-written parallel data
21
  - **ALMA-7B-LoRA**: Full-weight Fine-tune LLaMA-2-7B on 20B monolingual tokens and then **LoRA** fine-tune on human-written parallel data
22
+ - **ALMA-7B-R (NEW!)**: Further LoRA fine-tuning upon ALMA-7B-LoRA with contrastive preference optimization.
23
  - **ALMA-13B**: Full-weight Fine-tune LLaMA-2-7B on 12B monolingual tokens and then **Full-weight** fine-tune on human-written parallel data
24
  - **ALMA-13B-LoRA** (Our best system): Full-weight Fine-tune LLaMA-2-7B on 12B monolingual tokens and then **LoRA** fine-tune on human-written parallel data
25
+ - **ALMA-13B-R (NEW!)**: Further LoRA fine-tuning upon ALMA-13B-LoRA with contrastive preference optimization.
26
 
27
  Model checkpoints are released at huggingface:
28
  | Models | Base Model Link | LoRA Link |
29
  |:-------------:|:---------------:|:---------:|
30
  | ALMA-7B | [haoranxu/ALMA-7B](https://huggingface.co/haoranxu/ALMA-7B) | - |
31
  | ALMA-7B-LoRA | [haoranxu/ALMA-7B-Pretrain](https://huggingface.co/haoranxu/ALMA-7B-Pretrain) | [haoranxu/ALMA-7B-Pretrain-LoRA](https://huggingface.co/haoranxu/ALMA-7B-Pretrain-LoRA) |
32
+ | **ALMA-7B-R (NEW!)** | [haoranxu/ALMA-7B-Pretrain](https://huggingface.co/haoranxu/ALMA-7B-Pretrain) | [haoranxu/ALMA-7B-R](https://huggingface.co/haoranxu/ALMA-7B-R) |
33
  | ALMA-13B | [haoranxu/ALMA-13B](https://huggingface.co/haoranxu/ALMA-13B) | - |
34
  | ALMA-13B-LoRA | [haoranxu/ALMA-13B-Pretrain](https://huggingface.co/haoranxu/ALMA-13B-Pretrain) | [haoranxu/ALMA-13B-Pretrain-LoRA](https://huggingface.co/haoranxu/ALMA-13B-Pretrain-LoRA) |
35
+ | **ALMA-13B-R (NEW!)** | [haoranxu/ALMA-13B-Pretrain](https://huggingface.co/haoranxu/ALMA-13B-Pretrain) | [haoranxu/ALMA-13B-R](https://huggingface.co/haoranxu/ALMA-13B-R) |
36
 
37
+ **Note that `ALMA-7B-Pretrain` and `ALMA-13B-Pretrain` are NOT translation models. They only experience stage 1 monolingual fine-tuning (20B tokens for the 7B model and 12B tokens for the 13B model), and should be utilized in conjunction with their LoRA models.**
38
+
39
+ Datasets used by ALMA and ALMA-R are also released at huggingface now (NEW!)
40
+ | Datasets | Train / Validation| Test |
41
+ |:-------------:|:---------------:|:---------:|
42
+ | Human-Written Parallel Data (ALMA) | [train and validation](https://huggingface.co/datasets/haoranxu/ALMA-Human-Parallel) | [WMT'22](https://huggingface.co/datasets/haoranxu/WMT22-Test) |
43
+ | Triplet Preference Data | [train](https://huggingface.co/datasets/haoranxu/ALMA-R-Preference) | [WMT'22](https://huggingface.co/datasets/haoranxu/WMT22-Test) and [WMT'23](https://huggingface.co/datasets/haoranxu/WMT23-Test) |
44
 
45
  A quick start to use our best system (ALMA-13B-LoRA) for translation. An example of translating "我爱机器翻译。" into English:
46
  ```