|
--- |
|
license: gpl |
|
language: |
|
- en |
|
tags: |
|
- starcoder |
|
- wizardcoder |
|
- code |
|
- self-instruct |
|
- distillation |
|
--- |
|
|
|
# Model Card: Redmond-Hermes-Coder 15B |
|
|
|
## Model Description |
|
|
|
Redmond-Hermes-Coder 15B is a state-of-the-art language model fine-tuned on over 300,000 instructions. This model was fine-tuned by Nous Research, with Teknium and Karan4D leading the fine tuning process and dataset curation, Redmond AI sponsoring the compute, and several other contributors. |
|
|
|
This model was trained with a WizardCoder base, which itself uses a StarCoder base model. |
|
|
|
The model is truly great at code, but, it does come with a tradeoff though. While far better at code than the original Nous-Hermes built on Llama, it is worse than WizardCoder at pure code benchmarks, like HumanEval. |
|
|
|
It comes in at 39% on HumanEval, with WizardCoder at 57%. This is a preliminary experiment, and we are exploring improvements now. |
|
|
|
However, it does seem better at non-code than WizardCoder on a variety of things, including writing tasks. |
|
|
|
## Model Training |
|
|
|
The model was trained almost entirely on synthetic GPT-4 outputs. This includes data from diverse sources such as GPTeacher, the general, roleplay v1&2, code instruct datasets, Nous Instruct & PDACTL (unpublished), CodeAlpaca, Evol_Instruct Uncensored, GPT4-LLM, and Unnatural Instructions. |
|
|
|
Additional data inputs came from Camel-AI's Biology/Physics/Chemistry and Math Datasets, Airoboros' (v1) GPT-4 Dataset, and more from CodeAlpaca. The total volume of data encompassed over 300,000 instructions. |
|
|
|
## Collaborators |
|
The model fine-tuning and the datasets were a collaboration of efforts and resources from members of Nous Research, includingTeknium, Karan4D, Huemin Art, and Redmond AI's generous compute grants. |
|
|
|
Huge shoutout and acknowledgement is deserved for all the dataset creators who generously share their datasets openly. |
|
|
|
Among the contributors of datasets, GPTeacher was made available by Teknium, Wizard LM by nlpxucan, and the Nous Research Instruct Dataset was provided by Karan4D and HueminArt. |
|
The GPT4-LLM and Unnatural Instructions were provided by Microsoft, Airoboros dataset by jondurbin, Camel-AI datasets are from Camel-AI, and CodeAlpaca dataset by Sahil 2801. |
|
If anyone was left out, please open a thread in the community tab. |
|
|
|
## Prompt Format |
|
|
|
The model follows the Alpaca prompt format: |
|
``` |
|
### Instruction: |
|
|
|
### Response: |
|
``` |
|
|
|
or |
|
|
|
``` |
|
### Instruction: |
|
|
|
### Input: |
|
|
|
### Response: |
|
``` |
|
|
|
## Resources for Applied Use Cases: |
|
For an example of a back and forth chatbot using huggingface transformers and discord, check out: https://github.com/teknium1/alpaca-discord |
|
For an example of a roleplaying discord bot, check out this: https://github.com/teknium1/alpaca-roleplay-discordbot |
|
|
|
## Future Plans |
|
The model is currently being uploaded in FP16 format, and there are plans to convert the model to GGML and GPTQ 4bit quantizations. The team is also working on a full benchmark, similar to what was done for GPT4-x-Vicuna. We will try to get in discussions to get the model included in the GPT4All. |
|
|
|
## Benchmark Results |
|
``` |
|
HumanEval: 39% |
|
| Task |Version| Metric |Value | |Stderr| |
|
|------------------------------------------------|------:|---------------------|-----:|---|-----:| |
|
|arc_challenge | 0|acc |0.2858|± |0.0132| |
|
| | |acc_norm |0.3148|± |0.0136| |
|
|arc_easy | 0|acc |0.5349|± |0.0102| |
|
| | |acc_norm |0.5097|± |0.0103| |
|
|bigbench_causal_judgement | 0|multiple_choice_grade|0.5158|± |0.0364| |
|
|bigbench_date_understanding | 0|multiple_choice_grade|0.5230|± |0.0260| |
|
|bigbench_disambiguation_qa | 0|multiple_choice_grade|0.3295|± |0.0293| |
|
|bigbench_geometric_shapes | 0|multiple_choice_grade|0.1003|± |0.0159| |
|
| | |exact_str_match |0.0000|± |0.0000| |
|
|bigbench_logical_deduction_five_objects | 0|multiple_choice_grade|0.2260|± |0.0187| |
|
|bigbench_logical_deduction_seven_objects | 0|multiple_choice_grade|0.1957|± |0.0150| |
|
|bigbench_logical_deduction_three_objects | 0|multiple_choice_grade|0.3733|± |0.0280| |
|
|bigbench_movie_recommendation | 0|multiple_choice_grade|0.3200|± |0.0209| |
|
|bigbench_navigate | 0|multiple_choice_grade|0.4830|± |0.0158| |
|
|bigbench_reasoning_about_colored_objects | 0|multiple_choice_grade|0.4150|± |0.0110| |
|
|bigbench_ruin_names | 0|multiple_choice_grade|0.2143|± |0.0194| |
|
|bigbench_salient_translation_error_detection | 0|multiple_choice_grade|0.2926|± |0.0144| |
|
|bigbench_snarks | 0|multiple_choice_grade|0.5249|± |0.0372| |
|
|bigbench_sports_understanding | 0|multiple_choice_grade|0.4817|± |0.0159| |
|
|bigbench_temporal_sequences | 0|multiple_choice_grade|0.2700|± |0.0140| |
|
|bigbench_tracking_shuffled_objects_five_objects | 0|multiple_choice_grade|0.1864|± |0.0110| |
|
|bigbench_tracking_shuffled_objects_seven_objects| 0|multiple_choice_grade|0.1349|± |0.0082| |
|
|bigbench_tracking_shuffled_objects_three_objects| 0|multiple_choice_grade|0.3733|± |0.0280| |
|
|boolq | 1|acc |0.5498|± |0.0087| |
|
|hellaswag | 0|acc |0.3814|± |0.0048| |
|
| | |acc_norm |0.4677|± |0.0050| |
|
|openbookqa | 0|acc |0.1960|± |0.0178| |
|
| | |acc_norm |0.3100|± |0.0207| |
|
|piqa | 0|acc |0.6600|± |0.0111| |
|
| | |acc_norm |0.6610|± |0.0110| |
|
|winogrande | 0|acc |0.5343|± |0.0140| |
|
``` |
|
|
|
## Model Usage |
|
The model is available for download on Hugging Face. It is suitable for a wide range of language tasks, from generating creative text to understanding and following complex instructions. |
|
|
|
Compute provided by our project sponsor Redmond AI, thank you!! |
