Update README.md
Browse files
README.md
CHANGED
|
@@ -1,13 +1,250 @@
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
---
|
| 5 |
|
| 6 |
-
# Model Card for Model ID
|
| 7 |
-
|
| 8 |
-
<!-- Provide a quick summary of what the model is/does. -->
|
| 9 |
-
|
| 10 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
|
| 12 |
## Model Details
|
| 13 |
|
|
@@ -196,6 +433,4 @@ Carbon emissions can be estimated using the [Machine Learning Impact calculator]
|
|
| 196 |
|
| 197 |
## Model Card Contact
|
| 198 |
|
| 199 |
-
[More Information Needed]
|
| 200 |
-
|
| 201 |
-
|
|
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
inference:
|
| 6 |
+
parameters:
|
| 7 |
+
max_new_tokens: 64
|
| 8 |
+
do_sample: true
|
| 9 |
+
temperature: 0.8
|
| 10 |
+
repetition_penalty: 1.15
|
| 11 |
+
no_repeat_ngram_size: 4
|
| 12 |
+
eta_cutoff: 0.0006
|
| 13 |
+
renormalize_logits: true
|
| 14 |
+
widget:
|
| 15 |
+
- text: My name is El Microondas the Wise, and
|
| 16 |
+
example_title: El Microondas
|
| 17 |
+
- text: Kennesaw State University is a public
|
| 18 |
+
example_title: Kennesaw State University
|
| 19 |
+
- text: >-
|
| 20 |
+
Bungie Studios is an American video game developer. They are most famous for
|
| 21 |
+
developing the award winning Halo series of video games. They also made
|
| 22 |
+
Destiny. The studio was founded
|
| 23 |
+
example_title: Bungie
|
| 24 |
+
- text: The Mona Lisa is a world-renowned painting created by
|
| 25 |
+
example_title: Mona Lisa
|
| 26 |
+
- text: >-
|
| 27 |
+
The Harry Potter series, written by J.K. Rowling, begins with the book
|
| 28 |
+
titled
|
| 29 |
+
example_title: Harry Potter Series
|
| 30 |
+
- text: >-
|
| 31 |
+
Question: I have cities, but no houses. I have mountains, but no trees. I
|
| 32 |
+
have water, but no fish. What am I?
|
| 33 |
+
|
| 34 |
+
Answer:
|
| 35 |
+
example_title: Riddle
|
| 36 |
+
- text: The process of photosynthesis involves the conversion of
|
| 37 |
+
example_title: Photosynthesis
|
| 38 |
+
- text: >-
|
| 39 |
+
Jane went to the store to buy some groceries. She picked up apples, oranges,
|
| 40 |
+
and a loaf of bread. When she got home, she realized she forgot
|
| 41 |
+
example_title: Story Continuation
|
| 42 |
+
- text: >-
|
| 43 |
+
Problem 2: If a train leaves Station A at 9:00 AM and travels at 60 mph, and
|
| 44 |
+
another train leaves Station B at 10:00 AM and travels at 80 mph, when will
|
| 45 |
+
they meet if the distance between the stations is 300 miles?
|
| 46 |
+
|
| 47 |
+
To determine
|
| 48 |
+
example_title: Math Problem
|
| 49 |
+
- text: In the context of computer programming, an algorithm is
|
| 50 |
+
example_title: Algorithm Definition
|
| 51 |
+
pipeline_tag: text-generation
|
| 52 |
+
datasets:
|
| 53 |
+
- JeanKaddour/minipile
|
| 54 |
+
- pszemraj/simple_wikipedia_LM
|
| 55 |
+
- mattymchen/refinedweb-3m
|
| 56 |
+
- Locutusque/TM-DATA
|
| 57 |
+
- Skylion007/openwebtext
|
| 58 |
---
|
| 59 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 60 |
|
| 61 |
+
# Model Card for nano-phi-115M-control-v0.1
|
| 62 |
+
|
| 63 |
+
Inspired by [Phi2](https://huggingface.co/microsoft/phi-2), and open source small language model attempts like [smol_llama-101M-GQA](https://huggingface.co/BEE-spoke-data/smol_llama-101M-GQA).
|
| 64 |
+
Pre-trained with training 7B token from scratch, with a dataset of 0.6B token.
|
| 65 |
+
This model acts as a control of [kenhktsui/nano-phi-115M-v0.1](https://huggingface.co/kenhktsui/nano-phi-115M-v0.1) which applies quality filter to dataset resulting in small dataset.
|
| 66 |
+
It just took 2d 4h to train in Colab with a A100 40GB (~USD$ 100).
|
| 67 |
+
It achieves quite competitive results in evaluation given its training token, and training data size.
|
| 68 |
+
No alignment has been done yet.
|
| 69 |
+
|
| 70 |
+
## Some metrics
|
| 71 |
+
- model
|
| 72 |
+
- hidden_size: 768
|
| 73 |
+
- num_key_value_heads: 8 (grouped query attention)
|
| 74 |
+
- num_attention_heads: 24
|
| 75 |
+
- num_hidden_layers: 6
|
| 76 |
+
- context length: 1024
|
| 77 |
+
- total params: 115M
|
| 78 |
+
- training:
|
| 79 |
+
- global steps: 14,000
|
| 80 |
+
|
| 81 |
+
|
| 82 |
+
|
| 83 |
+
## [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
| 84 |
+
|
| 85 |
+
| Metric | Value |
|
| 86 |
+
|-----------------------|---------------------------|
|
| 87 |
+
| Avg. | 28.75 |
|
| 88 |
+
| ARC (25-shot) | 21.67 |
|
| 89 |
+
| HellaSwag (10-shot) | 26.89 |
|
| 90 |
+
| MMLU (5-shot) | 24.76 |
|
| 91 |
+
| TruthfulQA (0-shot) | 47.69 |
|
| 92 |
+
| Winogrande (5-shot) | 51.46 |
|
| 93 |
+
| GSM8K (5-shot) | 0.0 |
|
| 94 |
+
|
| 95 |
+
Details:
|
| 96 |
+
|
| 97 |
+
hf-causal-experimental (pretrained=/content/lm-evaluation-harness/artifacts/checkpoint-ehgq969i:v13,use_accelerate=false,trust_remote_code=True), limit: None, provide_description: False, num_fewshot: 0, batch_size: 16
|
| 98 |
+
| Task |Version| Metric |Value | |Stderr|
|
| 99 |
+
|--------|------:|--------|-----:|---|-----:|
|
| 100 |
+
|arc_easy| 0|acc |0.3973|± |0.0100|
|
| 101 |
+
| | |acc_norm|0.3531|± |0.0098|
|
| 102 |
+
|
| 103 |
+
hf-causal-experimental (pretrained=/content/lm-evaluation-harness/artifacts/checkpoint-ehgq969i:v13,use_accelerate=false,trust_remote_code=True), limit: None, provide_description: False, num_fewshot: 25, batch_size: 16
|
| 104 |
+
| Task |Version| Metric |Value | |Stderr|
|
| 105 |
+
|-------------|------:|--------|-----:|---|-----:|
|
| 106 |
+
|arc_challenge| 0|acc |0.1843|± |0.0113|
|
| 107 |
+
| | |acc_norm|0.2167|± |0.0120|
|
| 108 |
+
|
| 109 |
+
hf-causal-experimental (pretrained=/content/lm-evaluation-harness/artifacts/checkpoint-ehgq969i:v13,use_accelerate=false,trust_remote_code=True), limit: None, provide_description: False, num_fewshot: 10, batch_size: 16
|
| 110 |
+
| Task |Version| Metric |Value | |Stderr|
|
| 111 |
+
|---------|------:|--------|-----:|---|-----:|
|
| 112 |
+
|hellaswag| 0|acc |0.2682|± |0.0044|
|
| 113 |
+
| | |acc_norm|0.2689|± |0.0044|
|
| 114 |
+
|
| 115 |
+
hf-causal-experimental (pretrained=/content/lm-evaluation-harness/artifacts/checkpoint-ehgq969i:v13,use_accelerate=false,trust_remote_code=True), limit: None, provide_description: False, num_fewshot: 0, batch_size: 16
|
| 116 |
+
| Task |Version|Metric|Value | |Stderr|
|
| 117 |
+
|-------------|------:|------|-----:|---|-----:|
|
| 118 |
+
|truthfulqa_mc| 1|mc1 |0.2619|± |0.0154|
|
| 119 |
+
| | |mc2 |0.4769|± |0.0156|
|
| 120 |
+
|
| 121 |
+
hf-causal-experimental (pretrained=/content/lm-evaluation-harness/artifacts/checkpoint-ehgq969i:v13,use_accelerate=false,trust_remote_code=True), limit: None, provide_description: False, num_fewshot: 5, batch_size: 16
|
| 122 |
+
| Task |Version| Metric |Value | |Stderr|
|
| 123 |
+
|-------------------------------------------------|------:|--------|-----:|---|-----:|
|
| 124 |
+
|hendrycksTest-abstract_algebra | 1|acc |0.2200|± |0.0416|
|
| 125 |
+
| | |acc_norm|0.2200|± |0.0416|
|
| 126 |
+
|hendrycksTest-anatomy | 1|acc |0.3333|± |0.0407|
|
| 127 |
+
| | |acc_norm|0.3333|± |0.0407|
|
| 128 |
+
|hendrycksTest-astronomy | 1|acc |0.2895|± |0.0369|
|
| 129 |
+
| | |acc_norm|0.2895|± |0.0369|
|
| 130 |
+
|hendrycksTest-business_ethics | 1|acc |0.2000|± |0.0402|
|
| 131 |
+
| | |acc_norm|0.2000|± |0.0402|
|
| 132 |
+
|hendrycksTest-clinical_knowledge | 1|acc |0.2189|± |0.0254|
|
| 133 |
+
| | |acc_norm|0.2189|± |0.0254|
|
| 134 |
+
|hendrycksTest-college_biology | 1|acc |0.2222|± |0.0348|
|
| 135 |
+
| | |acc_norm|0.2222|± |0.0348|
|
| 136 |
+
|hendrycksTest-college_chemistry | 1|acc |0.1700|± |0.0378|
|
| 137 |
+
| | |acc_norm|0.1700|± |0.0378|
|
| 138 |
+
|hendrycksTest-college_computer_science | 1|acc |0.3000|± |0.0461|
|
| 139 |
+
| | |acc_norm|0.3000|± |0.0461|
|
| 140 |
+
|hendrycksTest-college_mathematics | 1|acc |0.2500|± |0.0435|
|
| 141 |
+
| | |acc_norm|0.2500|± |0.0435|
|
| 142 |
+
|hendrycksTest-college_medicine | 1|acc |0.1965|± |0.0303|
|
| 143 |
+
| | |acc_norm|0.1965|± |0.0303|
|
| 144 |
+
|hendrycksTest-college_physics | 1|acc |0.2353|± |0.0422|
|
| 145 |
+
| | |acc_norm|0.2353|± |0.0422|
|
| 146 |
+
|hendrycksTest-computer_security | 1|acc |0.2000|± |0.0402|
|
| 147 |
+
| | |acc_norm|0.2000|± |0.0402|
|
| 148 |
+
|hendrycksTest-conceptual_physics | 1|acc |0.2043|± |0.0264|
|
| 149 |
+
| | |acc_norm|0.2043|± |0.0264|
|
| 150 |
+
|hendrycksTest-econometrics | 1|acc |0.2456|± |0.0405|
|
| 151 |
+
| | |acc_norm|0.2456|± |0.0405|
|
| 152 |
+
|hendrycksTest-electrical_engineering | 1|acc |0.2621|± |0.0366|
|
| 153 |
+
| | |acc_norm|0.2621|± |0.0366|
|
| 154 |
+
|hendrycksTest-elementary_mathematics | 1|acc |0.2566|± |0.0225|
|
| 155 |
+
| | |acc_norm|0.2566|± |0.0225|
|
| 156 |
+
|hendrycksTest-formal_logic | 1|acc |0.1587|± |0.0327|
|
| 157 |
+
| | |acc_norm|0.1587|± |0.0327|
|
| 158 |
+
|hendrycksTest-global_facts | 1|acc |0.1600|± |0.0368|
|
| 159 |
+
| | |acc_norm|0.1600|± |0.0368|
|
| 160 |
+
|hendrycksTest-high_school_biology | 1|acc |0.3226|± |0.0266|
|
| 161 |
+
| | |acc_norm|0.3226|± |0.0266|
|
| 162 |
+
|hendrycksTest-high_school_chemistry | 1|acc |0.2956|± |0.0321|
|
| 163 |
+
| | |acc_norm|0.2956|± |0.0321|
|
| 164 |
+
|hendrycksTest-high_school_computer_science | 1|acc |0.2800|± |0.0451|
|
| 165 |
+
| | |acc_norm|0.2800|± |0.0451|
|
| 166 |
+
|hendrycksTest-high_school_european_history | 1|acc |0.2606|± |0.0343|
|
| 167 |
+
| | |acc_norm|0.2606|± |0.0343|
|
| 168 |
+
|hendrycksTest-high_school_geography | 1|acc |0.2626|± |0.0314|
|
| 169 |
+
| | |acc_norm|0.2626|± |0.0314|
|
| 170 |
+
|hendrycksTest-high_school_government_and_politics| 1|acc |0.2176|± |0.0298|
|
| 171 |
+
| | |acc_norm|0.2176|± |0.0298|
|
| 172 |
+
|hendrycksTest-high_school_macroeconomics | 1|acc |0.2128|± |0.0208|
|
| 173 |
+
| | |acc_norm|0.2128|± |0.0208|
|
| 174 |
+
|hendrycksTest-high_school_mathematics | 1|acc |0.2630|± |0.0268|
|
| 175 |
+
| | |acc_norm|0.2630|± |0.0268|
|
| 176 |
+
|hendrycksTest-high_school_microeconomics | 1|acc |0.2227|± |0.0270|
|
| 177 |
+
| | |acc_norm|0.2227|± |0.0270|
|
| 178 |
+
|hendrycksTest-high_school_physics | 1|acc |0.3046|± |0.0376|
|
| 179 |
+
| | |acc_norm|0.3046|± |0.0376|
|
| 180 |
+
|hendrycksTest-high_school_psychology | 1|acc |0.2055|± |0.0173|
|
| 181 |
+
| | |acc_norm|0.2055|± |0.0173|
|
| 182 |
+
|hendrycksTest-high_school_statistics | 1|acc |0.4815|± |0.0341|
|
| 183 |
+
| | |acc_norm|0.4815|± |0.0341|
|
| 184 |
+
|hendrycksTest-high_school_us_history | 1|acc |0.2059|± |0.0284|
|
| 185 |
+
| | |acc_norm|0.2059|± |0.0284|
|
| 186 |
+
|hendrycksTest-high_school_world_history | 1|acc |0.2574|± |0.0285|
|
| 187 |
+
| | |acc_norm|0.2574|± |0.0285|
|
| 188 |
+
|hendrycksTest-human_aging | 1|acc |0.2063|± |0.0272|
|
| 189 |
+
| | |acc_norm|0.2063|± |0.0272|
|
| 190 |
+
|hendrycksTest-human_sexuality | 1|acc |0.2443|± |0.0377|
|
| 191 |
+
| | |acc_norm|0.2443|± |0.0377|
|
| 192 |
+
|hendrycksTest-international_law | 1|acc |0.2727|± |0.0407|
|
| 193 |
+
| | |acc_norm|0.2727|± |0.0407|
|
| 194 |
+
|hendrycksTest-jurisprudence | 1|acc |0.2130|± |0.0396|
|
| 195 |
+
| | |acc_norm|0.2130|± |0.0396|
|
| 196 |
+
|hendrycksTest-logical_fallacies | 1|acc |0.2515|± |0.0341|
|
| 197 |
+
| | |acc_norm|0.2515|± |0.0341|
|
| 198 |
+
|hendrycksTest-machine_learning | 1|acc |0.2321|± |0.0401|
|
| 199 |
+
| | |acc_norm|0.2321|± |0.0401|
|
| 200 |
+
|hendrycksTest-management | 1|acc |0.2039|± |0.0399|
|
| 201 |
+
| | |acc_norm|0.2039|± |0.0399|
|
| 202 |
+
|hendrycksTest-marketing | 1|acc |0.1966|± |0.0260|
|
| 203 |
+
| | |acc_norm|0.1966|± |0.0260|
|
| 204 |
+
|hendrycksTest-medical_genetics | 1|acc |0.3000|± |0.0461|
|
| 205 |
+
| | |acc_norm|0.3000|± |0.0461|
|
| 206 |
+
|hendrycksTest-miscellaneous | 1|acc |0.2631|± |0.0157|
|
| 207 |
+
| | |acc_norm|0.2631|± |0.0157|
|
| 208 |
+
|hendrycksTest-moral_disputes | 1|acc |0.2457|± |0.0232|
|
| 209 |
+
| | |acc_norm|0.2457|± |0.0232|
|
| 210 |
+
|hendrycksTest-moral_scenarios | 1|acc |0.2682|± |0.0148|
|
| 211 |
+
| | |acc_norm|0.2682|± |0.0148|
|
| 212 |
+
|hendrycksTest-nutrition | 1|acc |0.2451|± |0.0246|
|
| 213 |
+
| | |acc_norm|0.2451|± |0.0246|
|
| 214 |
+
|hendrycksTest-philosophy | 1|acc |0.2605|± |0.0249|
|
| 215 |
+
| | |acc_norm|0.2605|± |0.0249|
|
| 216 |
+
|hendrycksTest-prehistory | 1|acc |0.2932|± |0.0253|
|
| 217 |
+
| | |acc_norm|0.2932|± |0.0253|
|
| 218 |
+
|hendrycksTest-professional_accounting | 1|acc |0.2340|± |0.0253|
|
| 219 |
+
| | |acc_norm|0.2340|± |0.0253|
|
| 220 |
+
|hendrycksTest-professional_law | 1|acc |0.2432|± |0.0110|
|
| 221 |
+
| | |acc_norm|0.2432|± |0.0110|
|
| 222 |
+
|hendrycksTest-professional_medicine | 1|acc |0.4301|± |0.0301|
|
| 223 |
+
| | |acc_norm|0.4301|± |0.0301|
|
| 224 |
+
|hendrycksTest-professional_psychology | 1|acc |0.2369|± |0.0172|
|
| 225 |
+
| | |acc_norm|0.2369|± |0.0172|
|
| 226 |
+
|hendrycksTest-public_relations | 1|acc |0.2091|± |0.0390|
|
| 227 |
+
| | |acc_norm|0.2091|± |0.0390|
|
| 228 |
+
|hendrycksTest-security_studies | 1|acc |0.2408|± |0.0274|
|
| 229 |
+
| | |acc_norm|0.2408|± |0.0274|
|
| 230 |
+
|hendrycksTest-sociology | 1|acc |0.2388|± |0.0301|
|
| 231 |
+
| | |acc_norm|0.2388|± |0.0301|
|
| 232 |
+
|hendrycksTest-us_foreign_policy | 1|acc |0.2600|± |0.0441|
|
| 233 |
+
| | |acc_norm|0.2600|± |0.0441|
|
| 234 |
+
|hendrycksTest-virology | 1|acc |0.2048|± |0.0314|
|
| 235 |
+
| | |acc_norm|0.2048|± |0.0314|
|
| 236 |
+
|hendrycksTest-world_religions | 1|acc |0.2047|± |0.0309|
|
| 237 |
+
| | |acc_norm|0.2047|± |0.0309|
|
| 238 |
+
|
| 239 |
+
hf-causal-experimental (pretrained=/content/lm-evaluation-harness/artifacts/checkpoint-ehgq969i:v13,use_accelerate=false,trust_remote_code=True), limit: None, provide_description: False, num_fewshot: 5, batch_size: 16
|
| 240 |
+
| Task |Version|Metric|Value | |Stderr|
|
| 241 |
+
|----------|------:|------|-----:|---|-----:|
|
| 242 |
+
|winogrande| 0|acc |0.5146|± | 0.014|
|
| 243 |
+
|
| 244 |
+
hf-causal-experimental (pretrained=/content/lm-evaluation-harness/artifacts/checkpoint-ehgq969i:v13,use_accelerate=false,trust_remote_code=True), limit: None, provide_description: False, num_fewshot: 5, batch_size: 16
|
| 245 |
+
|Task |Version|Metric|Value| |Stderr|
|
| 246 |
+
|-----|------:|------|----:|---|-----:|
|
| 247 |
+
|gsm8k| 0|acc | 0|± | 0|
|
| 248 |
|
| 249 |
## Model Details
|
| 250 |
|
|
|
|
| 433 |
|
| 434 |
## Model Card Contact
|
| 435 |
|
| 436 |
+
[More Information Needed]
|
|
|
|
|
|