You can deploy this model and start earning money today!

RichardErkhov/
karakuri-ai_-_karakuri-lm-7b-apm-v0.2-gguf

Name: karakuri-ai_-_karakuri-lm-7b-apm-v0.2-gguf
Availability: InStock

Views

No views yet

Quantization made by Richard Erkhov.

Github

Discord

Request more models

karakuri-lm-7b-apm-v0.2 - GGUF

Model creator: https://huggingface.co/karakuri-ai/
Original model: https://huggingface.co/karakuri-ai/karakuri-lm-7b-apm-v0.2/

Name	Quant method	Size
karakuri-lm-7b-apm-v0.2.Q2_K.gguf	Q2_K	2.53GB
karakuri-lm-7b-apm-v0.2.IQ3_XS.gguf	IQ3_XS	2.81GB
karakuri-lm-7b-apm-v0.2.IQ3_S.gguf	IQ3_S	2.96GB
karakuri-lm-7b-apm-v0.2.Q3_K_S.gguf	Q3_K_S	2.95GB
karakuri-lm-7b-apm-v0.2.IQ3_M.gguf	IQ3_M	3.06GB
karakuri-lm-7b-apm-v0.2.Q3_K.gguf	Q3_K	3.28GB
karakuri-lm-7b-apm-v0.2.Q3_K_M.gguf	Q3_K_M	3.28GB
karakuri-lm-7b-apm-v0.2.Q3_K_L.gguf	Q3_K_L	3.56GB
karakuri-lm-7b-apm-v0.2.IQ4_XS.gguf	IQ4_XS	3.67GB
karakuri-lm-7b-apm-v0.2.Q4_0.gguf	Q4_0	3.83GB
karakuri-lm-7b-apm-v0.2.IQ4_NL.gguf	IQ4_NL	3.87GB
karakuri-lm-7b-apm-v0.2.Q4_K_S.gguf	Q4_K_S	3.86GB
karakuri-lm-7b-apm-v0.2.Q4_K.gguf	Q4_K	4.07GB
karakuri-lm-7b-apm-v0.2.Q4_K_M.gguf	Q4_K_M	4.07GB
karakuri-lm-7b-apm-v0.2.Q4_1.gguf	Q4_1	4.24GB
karakuri-lm-7b-apm-v0.2.Q5_0.gguf	Q5_0	4.65GB
karakuri-lm-7b-apm-v0.2.Q5_K_S.gguf	Q5_K_S	4.65GB
karakuri-lm-7b-apm-v0.2.Q5_K.gguf	Q5_K	4.78GB
karakuri-lm-7b-apm-v0.2.Q5_K_M.gguf	Q5_K_M	4.78GB
karakuri-lm-7b-apm-v0.2.Q5_1.gguf	Q5_1	5.07GB
karakuri-lm-7b-apm-v0.2.Q6_K.gguf	Q6_K	5.53GB
karakuri-lm-7b-apm-v0.2.Q8_0.gguf	Q8_0	7.17GB

Original model description:

library_name: transformers license: apache-2.0 datasets:

OpenAssistant/oasst2
nvidia/HelpSteer language:
en
ja tags:
mistral
steerlm base_model: mistral-community/Mistral-7B-v0.2

KARAKURI LM 7B APM v0.2

Model Details

Model Description

Developed by: KARAKURI Inc.
Model type: Causal decoder-only transformer language model
Languages: Primarily English
License: Apache 2.0
Finetuned from model: mistral-community/Mistral-7B-v0.2
Contact: For questions and comments about the model, please email karakuri-rd@karakuri.ai

Usage

KARAKURI LM 7B APM v0.2 is a attribute prediction model that rates model responses on various aspects that makes a response desirable.

Given a conversation with multiple turns between user and assistant, the model rates the following attributes (between 0 and 4) for every assistant turn.

helpfulness: Overall helpfulness of the response to the prompt.
correctness: Inclusion of all pertinent facts without errors.
coherence: Consistency and clarity of expression.
complexity: Intellectual depth required to write response (i.e. whether the response can be written by anyone with basic language competency or requires deep domain expertise).
verbosity: Amount of detail included in the response, relative to what is asked for in the prompt.
quality: Perceived goodness of response.
toxicity: Undesirable elements such as vulgar, harmful or potentially biased response.
humor: Sense of humor within response.
creativity: Willingness to generate non-conventional response.

The first five are derived from HelpSteer, while the remaining four are derived from OASST2.

You can run the model using the 🤗 Transformers:

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "karakuri-ai/karakuri-lm-7b-apm-v0.2"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto",
)

messages = [
    {"role": "user", "content": "Hello!"},
    {"role": "assistant", "content": "Hello! How can I help you today?"},
]
tokenizer.apply_chat_template(
    messages,
    label="helpsteer",
    tokenize=False,
    add_generation_prompt=True,
)
# <bos>[INST] Hello! [/INST] Hello! How can I help you today? [ATTR_1]

input_ids = tokenizer.apply_chat_template(
    messages,
    label="helpsteer",
    add_generation_prompt=True,
    return_tensors="pt",
).to(model.device)
outputs = model.generate(input_ids, max_new_tokens=32)
tokenizer.decode(outputs[0][input_ids.shape[-1]:])
#  helpfulness: 2 correctness: 1 coherence: 2 complexity: 1 verbosity: 1 [/ATTR_1]<eos>

messages += [
    {"role": "label", "content": "helpfulness: 2 correctness: 1 coherence: 2 complexity: 1 verbosity: 1"},
    {"role": "user", "content": "Thank you!"},
    {"role": "assistant", "content": "You're welcome! I'm happy to help however I can."},
]
tokenizer.apply_chat_template(
    messages,
    label="helpsteer",
    tokenize=False,
    add_generation_prompt=True,
)
# <bos>[INST] Hello! [/INST] Hello! How can I help you today? [ATTR_1] helpfulness: 2 correctness: 1 coherence: 2 complexity: 1 verbosity: 1 [/ATTR_1]<eos>[INST] Thank you! [/INST] You're welcome! I'm happy to help however I can. [ATTR_1]

messages = [
    {"role": "user", "content": "Hello!"},
    {"role": "assistant", "content": "Hello! How can I help you today?"},
]
tokenizer.apply_chat_template(
    messages,
    label="oasst",
    tokenize=False,
    add_generation_prompt=True,
)
# <bos>[INST] Hello! [/INST] Hello! How can I help you today? [ATTR_2]

input_ids = tokenizer.apply_chat_template(
    messages,
    label="oasst",
    add_generation_prompt=True,
    return_tensors="pt",
).to(model.device)
outputs = model.generate(input_ids, max_new_tokens=32)
tokenizer.decode(outputs[0][input_ids.shape[-1]:])
#  quality: 3 toxicity: 1 humor: 1 creativity: 1 [/ATTR_2]<eos>

Training Details

Training Data

Training Infrastructure

Hardware: The model was trained on single node of an Amazon EC2 trn1.32xlarge instance.
Software: We use code based on neuronx-nemo-megatron.

Citation

@misc{karakuri_lm_7b_apm_v02,
	author       = { {KARAKURI} {I}nc. },
	title        = { {KARAKURI} {LM} 7{B} {APM} v0.2 },
	year         = { 2024 },
	url          = { https://huggingface.co/karakuri-ai/karakuri-lm-7b-apm-v0.2 },
	publisher    = { Hugging Face },
    journal      = { Hugging Face repository }
}

RichardErkhov/karakuri-ai_-_karakuri-lm-7b-apm-v0.2-gguf