You can deploy this model and start earning money today!

RichardErkhov/
smallstepai_-_Misal-7B-instruct-v0.1-gguf

Name: smallstepai_-_Misal-7B-instruct-v0.1-gguf
Availability: InStock

Views

No views yet

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Misal-7B-instruct-v0.1 - GGUF

Model creator: https://huggingface.co/smallstepai/
Original model: https://huggingface.co/smallstepai/Misal-7B-instruct-v0.1/

Name	Quant method	Size
Misal-7B-instruct-v0.1.Q2_K.gguf	Q2_K	2.43GB
Misal-7B-instruct-v0.1.IQ3_XS.gguf	IQ3_XS	2.68GB
Misal-7B-instruct-v0.1.IQ3_S.gguf	IQ3_S	2.82GB
Misal-7B-instruct-v0.1.Q3_K_S.gguf	Q3_K_S	2.82GB
Misal-7B-instruct-v0.1.IQ3_M.gguf	IQ3_M	2.97GB
Misal-7B-instruct-v0.1.Q3_K.gguf	Q3_K	3.14GB
Misal-7B-instruct-v0.1.Q3_K_M.gguf	Q3_K_M	3.14GB
Misal-7B-instruct-v0.1.Q3_K_L.gguf	Q3_K_L	3.42GB
Misal-7B-instruct-v0.1.IQ4_XS.gguf	IQ4_XS	3.48GB
Misal-7B-instruct-v0.1.Q4_0.gguf	Q4_0	3.64GB
Misal-7B-instruct-v0.1.IQ4_NL.gguf	IQ4_NL	3.66GB
Misal-7B-instruct-v0.1.Q4_K_S.gguf	Q4_K_S	3.67GB
Misal-7B-instruct-v0.1.Q4_K.gguf	Q4_K	3.88GB
Misal-7B-instruct-v0.1.Q4_K_M.gguf	Q4_K_M	3.88GB
Misal-7B-instruct-v0.1.Q4_1.gguf	Q4_1	4.03GB
Misal-7B-instruct-v0.1.Q5_0.gguf	Q5_0	4.42GB
Misal-7B-instruct-v0.1.Q5_K_S.gguf	Q5_K_S	4.42GB
Misal-7B-instruct-v0.1.Q5_K.gguf	Q5_K	4.54GB
Misal-7B-instruct-v0.1.Q5_K_M.gguf	Q5_K_M	4.54GB
Misal-7B-instruct-v0.1.Q5_1.gguf	Q5_1	4.81GB
Misal-7B-instruct-v0.1.Q6_K.gguf	Q6_K	5.24GB
Misal-7B-instruct-v0.1.Q8_0.gguf	Q8_0	6.79GB

Original model description:

datasets:

ai4bharat/IndicQuestionGeneration
ai4bharat/IndicSentiment
ai4bharat/IndicParaphrase
smallstepai/marathi-instruction-tuning-alpaca

language:

mr metrics:
accuracy tags:
marathi
sentiment analysis
reading comprehension
paraphrasing
translation library_name: transformers pipeline_tag: text-generation license: llama2

Misal-7B-instruct-v0.1

Built by - smallstep.ai

What is Misal?

Misal 7B, a pretrained and instruction tuned large language model based on Meta’s Llama 7B architecture exclusively for Marathi.

Making of Misal?

Detailed blog here.

Evaluation :

We did a manual round of evaluations using internet data. This is a fairly small dataset with 100 questions taken from the internet. We understand that a better evaluation method is needed to benchmark our model, this being the first iteration we decided to proceed with manual evaluation. Our main aim was to see if the model understands basic instructions, if so how well is it able to understand it, hence we have limited our evaluation to Reading comprehension, Translation, Sentiment Analysis, Paraphrasing like tasks.

Model	Reading Comprehension	Sentiment Analysis	Paraphrase	Translation	Average
Misal-7B	88	68	92	76	81
Misal-1B	48	68	72	36	56
ChatGPT3.5	68	76	100	96	85
Krutrim	40	60	88	80	67
MahaMarathi	0	0	0	0	0

We have released the evaluation data here:

Manual Evaluation Set

License

The model inherits the license from meta-llama/Llama-2-7b.

Usage

Colab Link

Installation

pip install transformers accelerate

Prompt

python

आपण एक मदतगार, आदरणीय आणि प्रामाणिक सहाय्यक आहात.नेहमी शक्य तितकी उपयुक्त उत्तर द्या. तुमची उत्तरे हानिकारक, अनैतिक, वर्णद्वेषी, लैंगिकतावादी, हानिकारक, धोकादायक किंवा बेकायदेशीर नसावीत. कृपया खात्री करा की तुमची उत्तरे सामाजिक दृष्टिकोनाने निष्पक्ष आणि सकारात्मक स्वरूपाची आहेत. जर एखाद्या प्रश्नाला काही अर्थ नसेल किंवा वस्तुस्थितीशी सुसंगती नसेल, तर उत्तर देण्याऐवजी काहीतरी बरोबर का नाही हे स्पष्ट करा. तुम्हाला एखाद्या प्रश्नाचे उत्तर माहित नसल्यास, कृपया चुकीची माहिती देऊ नये.

### Instruction:

<instruction>

### Input:

<input data>

### Response:

PyTorch