99 lines
3.7 KiB
Markdown
99 lines
3.7 KiB
Markdown
|
---
|
||
|
|
base_model: BEE-spoke-data/smol_llama-220M-open_instruct
|
||
|
|
datasets:
|
||
|
|
- VMware/open-instruct
|
||
|
|
inference: false
|
||
|
|
license: apache-2.0
|
||
|
|
model_creator: BEE-spoke-data
|
||
|
|
model_name: smol_llama-220M-open_instruct
|
||
|
|
pipeline_tag: text-generation
|
||
|
|
quantized_by: afrideva
|
||
|
|
tags:
|
||
|
|
- gguf
|
||
|
|
- ggml
|
||
|
|
- quantized
|
||
|
|
- q2_k
|
||
|
|
- q3_k_m
|
||
|
|
- q4_k_m
|
||
|
|
- q5_k_m
|
||
|
|
- q6_k
|
||
|
|
- q8_0
|
||
|
|
widget:
|
||
|
|
- example_title: burritos
|
||
|
|
text: "Below is an instruction that describes a task, paired with an input that
|
||
|
|
provides further context. Write a response that appropriately completes the request.
|
||
|
|
\ \n \n### Instruction: \n \nWrite an ode to Chipotle burritos. \n \n###
|
||
|
|
Response: \n"
|
||
|
|
---
|
||
|
|
# BEE-spoke-data/smol_llama-220M-open_instruct-GGUF
|
||
|
|
|
||
|
|
Quantized GGUF model files for [smol_llama-220M-open_instruct](https://huggingface.co/BEE-spoke-data/smol_llama-220M-open_instruct) from [BEE-spoke-data](https://huggingface.co/BEE-spoke-data)
|
||
|
|
|
||
|
|
|
||
|
|
| Name | Quant method | Size |
|
||
|
|
| ---- | ---- | ---- |
|
||
|
|
| [smol_llama-220m-open_instruct.fp16.gguf](https://huggingface.co/afrideva/smol_llama-220M-open_instruct-GGUF/resolve/main/smol_llama-220m-open_instruct.fp16.gguf) | fp16 | 436.50 MB |
|
||
|
|
| [smol_llama-220m-open_instruct.q2_k.gguf](https://huggingface.co/afrideva/smol_llama-220M-open_instruct-GGUF/resolve/main/smol_llama-220m-open_instruct.q2_k.gguf) | q2_k | 94.43 MB |
|
||
|
|
| [smol_llama-220m-open_instruct.q3_k_m.gguf](https://huggingface.co/afrideva/smol_llama-220M-open_instruct-GGUF/resolve/main/smol_llama-220m-open_instruct.q3_k_m.gguf) | q3_k_m | 114.65 MB |
|
||
|
|
| [smol_llama-220m-open_instruct.q4_k_m.gguf](https://huggingface.co/afrideva/smol_llama-220M-open_instruct-GGUF/resolve/main/smol_llama-220m-open_instruct.q4_k_m.gguf) | q4_k_m | 137.58 MB |
|
||
|
|
| [smol_llama-220m-open_instruct.q5_k_m.gguf](https://huggingface.co/afrideva/smol_llama-220M-open_instruct-GGUF/resolve/main/smol_llama-220m-open_instruct.q5_k_m.gguf) | q5_k_m | 157.91 MB |
|
||
|
|
| [smol_llama-220m-open_instruct.q6_k.gguf](https://huggingface.co/afrideva/smol_llama-220M-open_instruct-GGUF/resolve/main/smol_llama-220m-open_instruct.q6_k.gguf) | q6_k | 179.52 MB |
|
||
|
|
| [smol_llama-220m-open_instruct.q8_0.gguf](https://huggingface.co/afrideva/smol_llama-220M-open_instruct-GGUF/resolve/main/smol_llama-220m-open_instruct.q8_0.gguf) | q8_0 | 232.28 MB |
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
## Original Model Card:
|
||
|
|
# BEE-spoke-data/smol_llama-220M-open_instruct
|
||
|
|
|
||
|
|
> Please note that this is an experiment, and the model has limitations because it is smol.
|
||
|
|
|
||
|
|
|
||
|
|
prompt format is alpaca.
|
||
|
|
|
||
|
|
|
||
|
|
```
|
||
|
|
Below is an instruction that describes a task, paired with an input that
|
||
|
|
provides further context. Write a response that appropriately completes
|
||
|
|
the request.
|
||
|
|
|
||
|
|
### Instruction:
|
||
|
|
|
||
|
|
How can I increase my meme production/output? Currently, I only create them in ancient babylonian which is time consuming.
|
||
|
|
|
||
|
|
### Response:
|
||
|
|
```
|
||
|
|
|
||
|
|
This was **not** trained using a separate 'inputs' field (as `VMware/open-instruct` doesn't use one).
|
||
|
|
|
||
|
|
|
||
|
|
## Example
|
||
|
|
|
||
|
|
Output on the text above ^. The inference API is set to sample with low temp so you should see (_at least slightly_) different generations each time.
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
Note that the inference API parameters used here are an initial educated guess, and may be updated over time:
|
||
|
|
|
||
|
|
```yml
|
||
|
|
inference:
|
||
|
|
parameters:
|
||
|
|
do_sample: true
|
||
|
|
renormalize_logits: true
|
||
|
|
temperature: 0.25
|
||
|
|
top_p: 0.95
|
||
|
|
top_k: 50
|
||
|
|
min_new_tokens: 2
|
||
|
|
max_new_tokens: 96
|
||
|
|
repetition_penalty: 1.04
|
||
|
|
no_repeat_ngram_size: 6
|
||
|
|
epsilon_cutoff: 0.0006
|
||
|
|
```
|
||
|
|
|
||
|
|
Feel free to experiment with the parameters using the model in Python and let us know if you have improved results with other params!
|
||
|
|
|
||
|
|
## Data
|
||
|
|
|
||
|
|
This was trained on `VMware/open-instruct` so do whatever you want, provided it falls under the base apache-2.0 license :)
|
||
|
|
|
||
|
|
---
|