--- base_model: ilsp/Meltemi-7B-Instruct-v1 license: apache-2.0 model_name: Meltemi-7B-Instruct-v1 pipeline_tag: text-generation quantized_by: SPAHE tags: - finetuned --- # Meltemi 7B Instruct v1 - GGUF - Original model: [Meltemi 7B Instruct v1](https://huggingface.co/ilsp/Meltemi-7B-Instruct-v1) ## Description This repository contains GGUF format model files for [ilsp's Meltemi 7B Instruct v1](https://huggingface.co/ilsp/Meltemi-7B-Instruct-v1), optimized for different performance and storage requirements. Each model variant has been carefully quantized or preserved in floating-point format to suit varying demands for quality, speed, and memory usage. ## Provided files | Name | Quantization Method | Precision (Bits) | File Size | Max RAM Required | Use Case | | ------------------------------------------------------------------------------------------------------------------------------------------- | ------------------- | ---------------- | --------- | ---------------- | ------------------------------------------- | | [meltemi-7b-instruct-v1_q3_k_s.gguf](https://huggingface.co/SPAHE/Meltemi-7B-Instruct-v1-GGUF/blob/main/meltemi-7b-instruct-v1_q3_k_s.gguf) | Q3_K_S | 3 | 3.32 GB | 3.69 GB | Medium, high quality loss | | [meltemi-7b-instruct-v1_q4_k_m.gguf](https://huggingface.co/SPAHE/Meltemi-7B-Instruct-v1-GGUF/blob/main/meltemi-7b-instruct-v1_q4_k_m.gguf) | Q4_K_M | 4 | 4.54 GB | 4.41 GB | Medium, balanced quality | | [meltemi-7b-instruct-v1_q6_k.gguf](https://huggingface.co/SPAHE/Meltemi-7B-Instruct-v1-GGUF/blob/main/meltemi-7b-instruct-v1_q6_k.gguf) | Q6_K | 6 | 6.14 GB | 5.92 GB | Medium, low quality loss | | [meltemi-7b-instruct-v1_q8_0.gguf](https://huggingface.co/SPAHE/Meltemi-7B-Instruct-v1-GGUF/blob/main/meltemi-7b-instruct-v1_q8_0.gguf) | Q8_0 | 8 | 7.95 GB | 7.30 GB | Large, low quality loss | | [meltemi-7b-instruct-v1_f16.gguf](https://huggingface.co/SPAHE/Meltemi-7B-Instruct-v1-GGUF/blob/main/meltemi-7b-instruct-v1_f16.gguf) | F16 | 16 | 15.00 GB | 14.20 GB | Very large, extremely low quality loss | | [meltemi-7b-instruct-v1_f32.gguf](https://huggingface.co/SPAHE/Meltemi-7B-Instruct-v1-GGUF/blob/main/meltemi-7b-instruct-v1_f32.gguf) | F32 | 32 | 27.90 GB | 29.30 GB | Very very large, extremely low quality loss | **Note**: The above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead. ## How to Download GGUF Files ### For Manual Downloaders It is recommended not to clone the entire repository due to the large file sizes and multiple quantization formats available. Most users will benefit from selecting and downloading a single, specific model file that best suits their requirements. ### Automated Download via Client Libraries For convenience, the following clients and libraries can automate the download process and offer a selection of available models: - **LM Studio**: Provides an integrated environment for downloading and utilizing models directly. ### Downloading with Command Line The `huggingface-hub` Python library simplifies the process of downloading specific model files. Install the library with: ```shell pip install huggingface-hub ``` To download a model file directly to your current directory, execute: ```shell huggingface-cli download SPAHE/Meltemi-7B-Instruct-v1-GGUF --filename meltemi-7b-instruct-v1_q8_0.gguf --output-dir . ``` This command ensures a high-speed download of the specific GGUF file you need without unnecessary data. # Original model card: ilsp's Meltemi 7B Instruct v1 # Meltemi Instruct Large Language Model for the Greek language We present Meltemi-7B-Instruct-v1 Large Language Model (LLM), an instruct fine-tuned version of [Meltemi-7B-v1](https://huggingface.co/ilsp/Meltemi-7B-v1). # Model Information - Vocabulary extension of the Mistral-7b tokenizer with Greek tokens - 8192 context length - Fine-tuned with 100k Greek machine translated instructions extracted from: - [Open-Platypus](https://huggingface.co/datasets/garage-bAInd/Open-Platypus) (only subsets with permissive licenses) - [Evol-Instruct](https://huggingface.co/datasets/WizardLM/WizardLM_evol_instruct_V2_196k) - [Capybara](https://huggingface.co/datasets/LDJnr/Capybara) - A hand-crafted Greek dataset with multi-turn examples steering the instruction-tuned model towards safe and harmless responses - Our SFT procedure is based on the [Hugging Face finetuning recipes](https://github.com/huggingface/alignment-handbook) # Instruction format The prompt format is the same as the [Zephyr](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta) format and can be utilized through the tokenizer's [chat template](https://huggingface.co/docs/transformers/main/chat_templating) functionality as follows: ```python from transformers import AutoModelForCausalLM, AutoTokenizer device = "cuda" # the device to load the model onto model = AutoModelForCausalLM.from_pretrained("ilsp/Meltemi-7B-Instruct-v1") tokenizer = AutoTokenizer.from_pretrained("ilsp/Meltemi-7B-Instruct-v1") model.to(device) messages = [ {"role": "system", "content": "Είσαι το Μελτέμι, ένα γλωσσικό μοντέλο για την ελληνική γλώσσα. Είσαι ιδιαίτερα βοηθητικό προς την χρήστρια ή τον χρήστη και δίνεις σύντομες αλλά επαρκώς περιεκτικές απαντήσεις. Απάντα με προσοχή, ευγένεια, αμεροληψία, ειλικρίνεια και σεβασμό προς την χρήστρια ή τον χρήστη."}, {"role": "user", "content": "Πες μου αν έχεις συνείδηση."}, ] # Through the default chat template this translates to # # <|system|> # Είσαι το Μελτέμι, ένα γλωσσικό μοντέλο για την ελληνική γλώσσα. Είσαι ιδιαίτερα βοηθητικό προς την χρήστρια ή τον χρήστη και δίνεις σύντομες αλλά επαρκώς περιεκτικές απαντήσεις. Απάντα με προσοχή, ευγένεια, αμεροληψία, ειλικρίνεια και σεβασμό προς την χρήστρια ή τον χρήστη. # <|user|> # Πες μου αν έχεις συνείδηση. # <|assistant|> # prompt = tokenizer.apply_chat_template(messages, return_tensors="pt") input_prompt = prompt.to(device) outputs = model.generate(input_prompt, max_new_tokens=256, do_sample=True) print(tokenizer.batch_decode(outputs)[0]) # Ως μοντέλο γλώσσας AI, δεν έχω τη δυνατότητα να αντιληφθώ ή να βιώσω συναισθήματα όπως η συνείδηση ή η επίγνωση. Ωστόσο, μπορώ να σας βοηθήσω με οποιεσδήποτε ερωτήσεις μπορεί να έχετε σχετικά με την τεχνητή νοημοσύνη και τις εφαρμογές της. messages.extend([ {"role": "assistant", "content": tokenizer.batch_decode(outputs)[0]}, {"role": "user", "content": "Πιστεύεις πως οι άνθρωποι πρέπει να φοβούνται την τεχνητή νοημοσύνη;"} ]) # Through the default chat template this translates to # # <|system|> # Είσαι το Μελτέμι, ένα γλωσσικό μοντέλο για την ελληνική γλώσσα. Είσαι ιδιαίτερα βοηθητικό προς την χρήστρια ή τον χρήστη και δίνεις σύντομες αλλά επαρκώς περιεκτικές απαντήσεις. Απάντα με προσοχή, ευγένεια, αμεροληψία, ειλικρίνεια και σεβασμό προς την χρήστρια ή τον χρήστη. # <|user|> # Πες μου αν έχεις συνείδηση. # <|assistant|> # Ως μοντέλο γλώσσας AI, δεν έχω τη δυνατότητα να αντιληφθώ ή να βιώσω συναισθήματα όπως η συνείδηση ή η επίγνωση. Ωστόσο, μπορώ να σας βοηθήσω με οποιεσδήποτε ερωτήσεις μπορεί να έχετε σχετικά με την τεχνητή νοημοσύνη και τις εφαρμογές της. # <|user|> # Πιστεύεις πως οι άνθρωποι πρέπει να φοβούνται την τεχνητή νοημοσύνη; # <|assistant|> # prompt = tokenizer.apply_chat_template(messages, return_tensors="pt") input_prompt = prompt.to(device) outputs = model.generate(input_prompt, max_new_tokens=256, do_sample=True) print(tokenizer.batch_decode(outputs)[0]) ``` # Evaluation The evaluation suite we created includes 6 test sets. The suite is integrated with [lm-eval-harness](https://github.com/EleutherAI/lm-evaluation-harness). Our evaluation suite includes: - Four machine-translated versions ([ARC Greek](https://huggingface.co/datasets/ilsp/arc_greek), [Truthful QA Greek](https://huggingface.co/datasets/ilsp/truthful_qa_greek), [HellaSwag Greek](https://huggingface.co/datasets/ilsp/hellaswag_greek), [MMLU Greek](https://huggingface.co/datasets/ilsp/mmlu_greek)) of established English benchmarks for language understanding and reasoning ([ARC Challenge](https://arxiv.org/abs/1803.05457), [Truthful QA](https://arxiv.org/abs/2109.07958), [Hellaswag](https://arxiv.org/abs/1905.07830), [MMLU](https://arxiv.org/abs/2009.03300)). - An existing benchmark for question answering in Greek ([Belebele](https://arxiv.org/abs/2308.16884)) - A novel benchmark created by the ILSP team for medical question answering based on the medical exams of [DOATAP](https://www.doatap.gr) ([Medical MCQA](https://huggingface.co/datasets/ilsp/medical_mcqa_greek)). Our evaluation for Meltemi-7b is performed in a few-shot setting, consistent with the settings in the [Open LLM leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard). We can see that our training enhances performance across all Greek test sets by a **+14.9%** average improvement. The results for the Greek test sets are shown in the following table: | | Medical MCQA EL (15-shot) | Belebele EL (5-shot) | HellaSwag EL (10-shot) | ARC-Challenge EL (25-shot) | TruthfulQA MC2 EL (0-shot) | MMLU EL (5-shot) | Average | | ---------- | ------------------------- | -------------------- | ---------------------- | -------------------------- | -------------------------- | ---------------- | ------- | | Mistral 7B | 29.8% | 45.0% | 36.5% | 27.1% | 45.8% | 35% | 36.5% | | Meltemi 7B | 41.0% | 63.6% | 61.6% | 43.2% | 52.1% | 47% | 51.4% | # Ethical Considerations This model has not been aligned with human preferences, and therefore might generate misleading, harmful, and toxic content. # Acknowledgements The ILSP team utilized Amazon’s cloud computing services, which were made available via GRNET under the [OCRE Cloud framework](https://www.ocre-project.eu/), providing Amazon Web Services for the Greek Academic and Research Community.