--- 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 repo contains GGUF format model files for [ilsp's Meltemi 7B Instruct v1](https://huggingface.co/ilsp/Meltemi-7B-Instruct-v1). ### About GGUF GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp. Here is an incomplete list of clients and libraries that are known to support GGUF: * [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option. * [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration. * [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling. * [GPT4All](https://gpt4all.io/index.html), a free and open source local running GUI, supporting Windows, Linux and macOS with full GPU accel. * [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration. Linux available, in beta as of 27/11/2023. * [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection. * [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration. * [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server. * [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use. * [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server. Note, as of time of writing (November 27th 2023), ctransformers has not been updated in a long time and does not support many recent models. ## Compatibility These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) ## Provided files | Name | Quant method | Bits/Floats | Size | Max RAM required | Use case | | ---- | ---- | ---- | ---- | ---- | ----- | | [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 | 5 | 7.40 GB| 7.30 GB | very low quality loss - recommended | | [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 | 13.90 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 | 27.80 GB| 29.30 GB | very large, extremely low quality loss - not recommended | **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 **Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file. The following clients/libraries will automatically download models for you, providing a list of available models to choose from: * LM Studio * LoLLMS Web UI * Faraday.dev ### On the command line, including multiple files at once I recommend using the `huggingface-hub` Python library: ```shell pip3 install huggingface-hub ``` Then you can download any individual model file to the current directory, at high speed, with a command like this: ```shell huggingface-cli download SPAHE/Meltemi-7B-Instruct-v1-GGUF meltemi-7B-instruct-v1_q8_0.gguf --local-dir . --local-dir-use-symlinks False ``` # 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.