--- license: cc-by-nc-4.0 language: - ro base_model: - OpenLLM-Ro/RoLlama2-7b-Instruct-2024-10-09 datasets: - OpenLLM-Ro/ro_dpo_helpsteer model-index: - name: OpenLLM-Ro/RoLlama2-7b-Instruct-DPO-2024-10-09 results: - task: type: text-generation dataset: name: RoMT-Bench type: RoMT-Bench metrics: - name: Score type: Score value: 4.61 - task: type: text-generation dataset: name: RoCulturaBench type: RoCulturaBench metrics: - name: Score type: Score value: 4.80 - task: type: text-generation dataset: name: Romanian_Academic_Benchmarks type: Romanian_Academic_Benchmarks metrics: - name: Average accuracy type: accuracy value: 43.20 - task: type: text-generation dataset: name: OpenLLM-Ro/ro_arc_challenge type: OpenLLM-Ro/ro_arc_challenge metrics: - name: Average accuracy type: accuracy value: 44.24 - task: type: text-generation dataset: name: OpenLLM-Ro/ro_mmlu type: OpenLLM-Ro/ro_mmlu metrics: - name: Average accuracy type: accuracy value: 38.39 - task: type: text-generation dataset: name: OpenLLM-Ro/ro_winogrande type: OpenLLM-Ro/ro_winogrande metrics: - name: Average accuracy type: accuracy value: 62.57 - task: type: text-generation dataset: name: OpenLLM-Ro/ro_hellaswag type: OpenLLM-Ro/ro_hellaswag metrics: - name: Average accuracy type: accuracy value: 59.20 - task: type: text-generation dataset: name: OpenLLM-Ro/ro_gsm8k type: OpenLLM-Ro/ro_gsm8k metrics: - name: Average accuracy type: accuracy value: 15.72 - task: type: text-generation dataset: name: OpenLLM-Ro/ro_truthfulqa type: OpenLLM-Ro/ro_truthfulqa metrics: - name: Average accuracy type: accuracy value: 39.07 - task: type: text-generation dataset: name: LaRoSeDa_binary type: LaRoSeDa_binary metrics: - name: Average macro-f1 type: macro-f1 value: 97.31 - task: type: text-generation dataset: name: LaRoSeDa_multiclass type: LaRoSeDa_multiclass metrics: - name: Average macro-f1 type: macro-f1 value: 60.56 - task: type: text-generation dataset: name: LaRoSeDa_binary_finetuned type: LaRoSeDa_binary_finetuned metrics: - name: Average macro-f1 type: macro-f1 value: 0.00 - task: type: text-generation dataset: name: LaRoSeDa_multiclass_finetuned type: LaRoSeDa_multiclass_finetuned metrics: - name: Average macro-f1 type: macro-f1 value: 0.00 - task: type: text-generation dataset: name: WMT_EN-RO type: WMT_EN-RO metrics: - name: Average bleu type: bleu value: 26.56 - task: type: text-generation dataset: name: WMT_RO-EN type: WMT_RO-EN metrics: - name: Average bleu type: bleu value: 21.68 - task: type: text-generation dataset: name: WMT_EN-RO_finetuned type: WMT_EN-RO_finetuned metrics: - name: Average bleu type: bleu value: 0.00 - task: type: text-generation dataset: name: WMT_RO-EN_finetuned type: WMT_RO-EN_finetuned metrics: - name: Average bleu type: bleu value: 0.00 - task: type: text-generation dataset: name: XQuAD type: XQuAD metrics: - name: Average exact_match type: exact_match value: 35.78 - task: type: text-generation dataset: name: XQuAD type: XQuAD metrics: - name: Average f1 type: f1 value: 59.31 - task: type: text-generation dataset: name: XQuAD_finetuned type: XQuAD_finetuned metrics: - name: Average exact_match type: exact_match value: 0.00 - task: type: text-generation dataset: name: XQuAD_finetuned type: XQuAD_finetuned metrics: - name: Average f1 type: f1 value: 0.00 - task: type: text-generation dataset: name: STS type: STS metrics: - name: Average spearman type: spearman value: 61.22 - task: type: text-generation dataset: name: STS type: STS metrics: - name: Average pearson type: pearson value: 58.41 - task: type: text-generation dataset: name: STS_finetuned type: STS_finetuned metrics: - name: Average spearman type: spearman value: 0.00 - task: type: text-generation dataset: name: STS_finetuned type: STS_finetuned metrics: - name: Average pearson type: pearson value: 0.00 - task: type: text-generation dataset: name: RoMT-Bench type: RoMT-Bench metrics: - name: First turn type: Score value: 5.15 - name: Second turn type: Score value: 4.06 - task: type: text-generation dataset: name: OpenLLM-Ro/ro_arc_challenge type: OpenLLM-Ro/ro_arc_challenge metrics: - name: 0-shot type: accuracy value: 42.67 - name: 1-shot type: accuracy value: 43.36 - name: 3-shot type: accuracy value: 44.13 - name: 5-shot type: accuracy value: 44.30 - name: 10-shot type: accuracy value: 45.67 - name: 25-shot type: accuracy value: 45.33 - task: type: text-generation dataset: name: OpenLLM-Ro/ro_mmlu type: OpenLLM-Ro/ro_mmlu metrics: - name: 0-shot type: accuracy value: 36.62 - name: 1-shot type: accuracy value: 38.04 - name: 3-shot type: accuracy value: 39.52 - name: 5-shot type: accuracy value: 39.36 - task: type: text-generation dataset: name: OpenLLM-Ro/ro_winogrande type: OpenLLM-Ro/ro_winogrande metrics: - name: 0-shot type: accuracy value: 61.72 - name: 1-shot type: accuracy value: 62.04 - name: 3-shot type: accuracy value: 63.85 - name: 5-shot type: accuracy value: 62.67 - task: type: text-generation dataset: name: OpenLLM-Ro/ro_hellaswag type: OpenLLM-Ro/ro_hellaswag metrics: - name: 0-shot type: accuracy value: 58.75 - name: 1-shot type: accuracy value: 58.29 - name: 3-shot type: accuracy value: 59.28 - name: 5-shot type: accuracy value: 59.68 - name: 10-shot type: accuracy value: 60.01 - task: type: text-generation dataset: name: OpenLLM-Ro/ro_gsm8k type: OpenLLM-Ro/ro_gsm8k metrics: - name: 1-shot type: accuracy value: 11.14 - name: 3-shot type: accuracy value: 17.97 - name: 5-shot type: accuracy value: 18.04 - task: type: text-generation dataset: name: LaRoSeDa_binary type: LaRoSeDa_binary metrics: - name: 0-shot type: macro-f1 value: 98.03 - name: 1-shot type: macro-f1 value: 95.96 - name: 3-shot type: macro-f1 value: 97.33 - name: 5-shot type: macro-f1 value: 97.90 - task: type: text-generation dataset: name: LaRoSeDa_multiclass type: LaRoSeDa_multiclass metrics: - name: 0-shot type: macro-f1 value: 60.67 - name: 1-shot type: macro-f1 value: 51.37 - name: 3-shot type: macro-f1 value: 62.49 - name: 5-shot type: macro-f1 value: 67.70 - task: type: text-generation dataset: name: WMT_EN-RO type: WMT_EN-RO metrics: - name: 0-shot type: bleu value: 19.83 - name: 1-shot type: bleu value: 29.04 - name: 3-shot type: bleu value: 28.90 - name: 5-shot type: bleu value: 28.47 - task: type: text-generation dataset: name: WMT_RO-EN type: WMT_RO-EN metrics: - name: 0-shot type: bleu value: 1.74 - name: 1-shot type: bleu value: 15.28 - name: 3-shot type: bleu value: 34.13 - name: 5-shot type: bleu value: 35.56 - task: type: text-generation dataset: name: XQuAD_EM type: XQuAD_EM metrics: - name: 0-shot type: exact_match value: 26.97 - name: 1-shot type: exact_match value: 36.30 - name: 3-shot type: exact_match value: 40.25 - name: 5-shot type: exact_match value: 39.58 - task: type: text-generation dataset: name: XQuAD_F1 type: XQuAD_F1 metrics: - name: 0-shot type: f1 value: 52.90 - name: 1-shot type: f1 value: 60.05 - name: 3-shot type: f1 value: 62.08 - name: 5-shot type: f1 value: 62.22 - task: type: text-generation dataset: name: STS_Spearman type: STS_Spearman metrics: - name: 1-shot type: spearman value: 62.07 - name: 3-shot type: spearman value: 59.47 - name: 5-shot type: spearman value: 62.12 - task: type: text-generation dataset: name: STS_Pearson type: STS_Pearson metrics: - name: 1-shot type: pearson value: 60.60 - name: 3-shot type: pearson value: 56.44 - name: 5-shot type: pearson value: 58.18 --- # Model Card for Model ID This model points/is identical to [RoLlama2-7b-Instruct-DPO-2024-10-09](https://huggingface.co/OpenLLM-Ro/RoLlama2-7b-Instruct-DPO-2024-10-09). RoLlama2 is a family of pretrained and fine-tuned generative text models for Romanian. This is the repository for the **human aligned instruct 7B model**. Links to other models can be found at the bottom of this page. ## Model Details ### Model Description OpenLLM represents the first open-source effort to build a LLM specialized for Romanian. OpenLLM-Ro developed and publicly releases a collection of Romanian LLMs, both in the form of foundational model and instruct and chat variants. - **Developed by:** OpenLLM-Ro - **Language(s):** Romanian - **License:** cc-by-nc-4.0 - **Finetuned from model:** [RoLlama2-7b-Instruct-2024-10-09](https://huggingface.co/OpenLLM-Ro/RoLlama2-7b-Instruct-2024-10-09) - **Trained using:** [RoHelpSteer](https://huggingface.co/datasets/OpenLLM-Ro/ro_dpo_helpsteer) ### Model Sources - **Repository:** https://github.com/OpenLLM-Ro/LLaMA-Factory - **Paper:** https://arxiv.org/abs/2406.18266 ## Intended Use ### Intended Use Cases RoLlama2 is intented for research use in Romanian. Base models can be adapted for a variety of natural language tasks while instruction and chat tuned models are intended for assistant-like chat. ### Out-of-Scope Use Use in any manner that violates the license, any applicable laws or regluations, use in languages other than Romanian. ## How to Get Started with the Model Use the code below to get started with the model. ```python from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("OpenLLM-Ro/RoLlama2-7b-Instruct-DPO") model = AutoModelForCausalLM.from_pretrained("OpenLLM-Ro/RoLlama2-7b-Instruct-DPO") instruction = "Care este cel mai înalt vârf muntos din România?" chat = [ {"role": "system", "content": "Ești un asistent folositor, respectuos și onest. Încearcă să ajuți cât mai mult prin informațiile oferite, excluzând răspunsuri toxice, rasiste, sexiste, periculoase și ilegale."}, {"role": "user", "content": instruction}, ] prompt = tokenizer.apply_chat_template(chat, tokenize=False) inputs = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt") outputs = model.generate(input_ids=inputs, max_new_tokens=128) print(tokenizer.decode(outputs[0])) ``` ## Academic Benchmarks
Model
Average
ARC
MMLU
Winogrande
Hellaswag
GSM8k
TruthfulQA
Llama-2-7b-chat
36.84
37.03
33.80
55.87
45.36
4.90
44.09
RoLlama2-7b-Instruct-2024-05-14
45.71
43.66
39.70
70.34
57.36
18.78
44.44
RoLlama2-7b-Instruct-2024-10-09
44.50
44.73
40.39
63.67
59.12
13.29
45.78
RoLlama2-7b-Instruct-DPO-2024-10-09
43.20
44.24
38.39
62.57
59.20
15.72
39.07
## Downstream tasks <
LaRoSeDa
WMT
Few-shot
Finetuned
Few-shot
Finetuned
Model
Binary
(Macro F1)
Multiclass
(Macro F1)
Binary
(Macro F1)
Multiclass
(Macro F1)
EN-RO
(Bleu)
RO-EN
(Bleu)
EN-RO
(Bleu)
RO-EN
(Bleu)
Llama-2-7b-chat
87.78
52.81
97.27
82.02
15.55
28.53
19.99
31.48
RoLlama2-7b-Instruct-2024-05-14
97.48
65.26
98.83
87.28
27.38
10.32
27.59
40.13
RoLlama2-7b-Instruct-2024-10-09
97.66
62.41
97.97
60.89
27.13
19.39
27.63
39.75
RoLlama2-7b-Instruct-DPO-2024-10-09
97.31
60.56
-
-
26.56
21.68
-
-
XQuAD
STS
Few-shot
Finetuned
Few-shot
Finetuned
Model
(EM)
(F1)
(EM)
(F1)
(Spearman)
(Pearson)
(Spearman)
(Pearson)
Llama-2-7b-chat
32.35
54.00
60.34
75.98
32.56
31.99
74.08
72.64
RoLlama2-7b-Instruct-2024-05-14
44.52
64.75
54.96
70.20
65.50
67.79
84.44
84.76
RoLlama2-7b-Instruct-2024-10-09
45.71
65.08
59.24
74.25
59.69
57.16
84.66
85.07
RoLlama2-7b-Instruct-DPO-2024-10-09
35.78
59.31
-
-
61.22
58.41
-
-
## Romanian MT-Bench
Model
Average
1st turn
2nd turn
Answers in Ro
Llama-2-7b-chat
1.08
1.44
0.73
45/160
RoLlama2-7b-Instruct-2024-05-14
3.86
4.67
3.04
160/160
RoLlama2-7b-Instruct-2024-10-09
4.43
4.92
3.94
160/160
RoLlama2-7b-Instruct-DPO-2024-10-09
4.61
5.15
4.06
160/160
## RoCulturaBench
Model
Average
Answers in Ro
Llama-2-7b-chat
1.21
33/100
RoLlama2-7b-Instruct-2024-05-14
3.77
100/100
RoLlama2-7b-Instruct-2024-10-09
4.08
100/100
RoLlama2-7b-Instruct-DPO-2024-10-09
4.80
100/100
## RoLlama2 Model Family | Model | Link | |--------------------|:--------:| |RoLlama2-7b-Base-2024-05-14 | [link](https://huggingface.co/OpenLLM-Ro/RoLlama2-7b-Base-2024-05-14) | |RoLlama2-7b-Instruct-2024-05-14 | [link](https://huggingface.co/OpenLLM-Ro/RoLlama2-7b-Instruct-2024-05-14) | |RoLlama2-7b-Instruct-2024-10-09| [link](https://huggingface.co/OpenLLM-Ro/RoLlama2-7b-Instruct-2024-10-09) | |*RoLlama2-7b-Instruct-DPO-2024-10-09*| [link](https://huggingface.co/OpenLLM-Ro/RoLlama2-7b-Instruct-DPO-2024-10-09) | ## Citation ``` @misc{masala2024vorbecstiromanecsterecipetrain, title={"Vorbe\c{s}ti Rom\^ane\c{s}te?" A Recipe to Train Powerful Romanian LLMs with English Instructions}, author={Mihai Masala and Denis C. Ilie-Ablachim and Alexandru Dima and Dragos Corlatescu and Miruna Zavelca and Ovio Olaru and Simina Terian-Dan and Andrei Terian-Dan and Marius Leordeanu and Horia Velicu and Marius Popescu and Mihai Dascalu and Traian Rebedea}, year={2024}, eprint={2406.18266}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2406.18266}, } ```