New model from https://wandb.ai/wandb/huggingtweets/runs/2q7qp6sj
Browse files- README.md +37 -65
- config.json +7 -0
- merges.txt +1 -1
- pytorch_model.bin +2 -2
- special_tokens_map.json +1 -1
- tokenizer.json +0 -0
- tokenizer_config.json +1 -1
- flax_model.msgpack → training_args.bin +2 -2
- vocab.json +0 -0
README.md
CHANGED
|
@@ -1,28 +1,27 @@
|
|
| 1 |
---
|
| 2 |
language: en
|
| 3 |
-
thumbnail: https://
|
| 4 |
tags:
|
| 5 |
- huggingtweets
|
| 6 |
widget:
|
| 7 |
- text: "My dream is"
|
| 8 |
---
|
| 9 |
|
| 10 |
-
<
|
| 11 |
-
|
| 12 |
-
<
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
</
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
<div style="
|
| 23 |
-
</div>
|
| 24 |
-
<div style="
|
| 25 |
-
<div style="font-size: 15px; color: #657786">@magicrealismbot bot</div>
|
| 26 |
</div>
|
| 27 |
|
| 28 |
I was made with [huggingtweets](https://github.com/borisdayma/huggingtweets).
|
|
@@ -35,62 +34,41 @@ The model uses the following pipeline.
|
|
| 35 |
|
| 36 |

|
| 37 |
|
| 38 |
-
To understand how the model was developed, check the [W&B report](https://
|
| 39 |
|
| 40 |
## Training data
|
| 41 |
|
| 42 |
-
The model was trained on
|
| 43 |
-
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
<tr style='border-width:0 0 1px 0; border-color: #E2E8F0'>
|
| 53 |
-
<td style='border-width:0'>Tweets downloaded</td>
|
| 54 |
-
<td style='border-width:0'>3204</td>
|
| 55 |
-
</tr>
|
| 56 |
-
<tr style='border-width:0 0 1px 0; border-color: #E2E8F0'>
|
| 57 |
-
<td style='border-width:0'>Retweets</td>
|
| 58 |
-
<td style='border-width:0'>0</td>
|
| 59 |
-
</tr>
|
| 60 |
-
<tr style='border-width:0 0 1px 0; border-color: #E2E8F0'>
|
| 61 |
-
<td style='border-width:0'>Short tweets</td>
|
| 62 |
-
<td style='border-width:0'>0</td>
|
| 63 |
-
</tr>
|
| 64 |
-
<tr style='border-width:0'>
|
| 65 |
-
<td style='border-width:0'>Tweets kept</td>
|
| 66 |
-
<td style='border-width:0'>3204</td>
|
| 67 |
-
</tr>
|
| 68 |
-
</tbody>
|
| 69 |
-
</table>
|
| 70 |
-
|
| 71 |
-
[Explore the data](https://app.wandb.ai/wandb/huggingtweets/runs/1hpsx1pa/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
|
| 72 |
|
| 73 |
## Training procedure
|
| 74 |
|
| 75 |
The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on @magicrealismbot's tweets.
|
| 76 |
|
| 77 |
-
Hyperparameters and metrics are recorded in the [W&B training run](https://
|
| 78 |
-
|
| 79 |
-
At the end of training, [the final model](https://app.wandb.ai/wandb/huggingtweets/runs/c0z341v9/artifacts) is logged and versioned.
|
| 80 |
|
| 81 |
-
|
| 82 |
|
| 83 |
-
|
| 84 |
|
| 85 |
You can use this model directly with a pipeline for text generation:
|
| 86 |
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
|
| 90 |
-
|
| 91 |
-
|
|
|
|
| 92 |
|
| 93 |
-
|
| 94 |
|
| 95 |
The model suffers from [the same limitations and bias as GPT-2](https://huggingface.co/gpt2#limitations-and-bias).
|
| 96 |
|
|
@@ -100,14 +78,8 @@ In addition, the data present in the user's tweets further affects the text gene
|
|
| 100 |
|
| 101 |
*Built by Boris Dayma*
|
| 102 |
|
| 103 |
-
</section>
|
| 104 |
-
|
| 105 |
[](https://twitter.com/intent/follow?screen_name=borisdayma)
|
| 106 |
|
| 107 |
-
<section class='prose'>
|
| 108 |
For more details, visit the project repository.
|
| 109 |
-
</section>
|
| 110 |
|
| 111 |
[](https://github.com/borisdayma/huggingtweets)
|
| 112 |
-
|
| 113 |
-
<!--- random size file -->
|
|
|
|
| 1 |
---
|
| 2 |
language: en
|
| 3 |
+
thumbnail: https://github.com/borisdayma/huggingtweets/blob/master/img/logo.png?raw=true
|
| 4 |
tags:
|
| 5 |
- huggingtweets
|
| 6 |
widget:
|
| 7 |
- text: "My dream is"
|
| 8 |
---
|
| 9 |
|
| 10 |
+
<div class="inline-flex flex-col" style="line-height: 1.5;">
|
| 11 |
+
<div class="flex">
|
| 12 |
+
<div
|
| 13 |
+
style="display:inherit; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://pbs.twimg.com/profile_images/668872745329885184/67TNOs2A_400x400.png')">
|
| 14 |
+
</div>
|
| 15 |
+
<div
|
| 16 |
+
style="display:none; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('')">
|
| 17 |
+
</div>
|
| 18 |
+
<div
|
| 19 |
+
style="display:none; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('')">
|
| 20 |
+
</div>
|
| 21 |
+
</div>
|
| 22 |
+
<div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">🤖 AI BOT 🤖</div>
|
| 23 |
+
<div style="text-align: center; font-size: 16px; font-weight: 800">Magic Realism Bot</div>
|
| 24 |
+
<div style="text-align: center; font-size: 14px;">@magicrealismbot</div>
|
|
|
|
| 25 |
</div>
|
| 26 |
|
| 27 |
I was made with [huggingtweets](https://github.com/borisdayma/huggingtweets).
|
|
|
|
| 34 |
|
| 35 |

|
| 36 |
|
| 37 |
+
To understand how the model was developed, check the [W&B report](https://wandb.ai/wandb/huggingtweets/reports/HuggingTweets-Train-a-Model-to-Generate-Tweets--VmlldzoxMTY5MjI).
|
| 38 |
|
| 39 |
## Training data
|
| 40 |
|
| 41 |
+
The model was trained on tweets from Magic Realism Bot.
|
| 42 |
+
|
| 43 |
+
| Data | Magic Realism Bot |
|
| 44 |
+
| --- | --- |
|
| 45 |
+
| Tweets downloaded | 3250 |
|
| 46 |
+
| Retweets | 0 |
|
| 47 |
+
| Short tweets | 0 |
|
| 48 |
+
| Tweets kept | 3250 |
|
| 49 |
+
|
| 50 |
+
[Explore the data](https://wandb.ai/wandb/huggingtweets/runs/2saq2zag/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 51 |
|
| 52 |
## Training procedure
|
| 53 |
|
| 54 |
The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on @magicrealismbot's tweets.
|
| 55 |
|
| 56 |
+
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/wandb/huggingtweets/runs/2q7qp6sj) for full transparency and reproducibility.
|
|
|
|
|
|
|
| 57 |
|
| 58 |
+
At the end of training, [the final model](https://wandb.ai/wandb/huggingtweets/runs/2q7qp6sj/artifacts) is logged and versioned.
|
| 59 |
|
| 60 |
+
## How to use
|
| 61 |
|
| 62 |
You can use this model directly with a pipeline for text generation:
|
| 63 |
|
| 64 |
+
```python
|
| 65 |
+
from transformers import pipeline
|
| 66 |
+
generator = pipeline('text-generation',
|
| 67 |
+
model='huggingtweets/magicrealismbot')
|
| 68 |
+
generator("My dream is", num_return_sequences=5)
|
| 69 |
+
```
|
| 70 |
|
| 71 |
+
## Limitations and bias
|
| 72 |
|
| 73 |
The model suffers from [the same limitations and bias as GPT-2](https://huggingface.co/gpt2#limitations-and-bias).
|
| 74 |
|
|
|
|
| 78 |
|
| 79 |
*Built by Boris Dayma*
|
| 80 |
|
|
|
|
|
|
|
| 81 |
[](https://twitter.com/intent/follow?screen_name=borisdayma)
|
| 82 |
|
|
|
|
| 83 |
For more details, visit the project repository.
|
|
|
|
| 84 |
|
| 85 |
[](https://github.com/borisdayma/huggingtweets)
|
|
|
|
|
|
config.json
CHANGED
|
@@ -1,4 +1,5 @@
|
|
| 1 |
{
|
|
|
|
| 2 |
"activation_function": "gelu_new",
|
| 3 |
"architectures": [
|
| 4 |
"GPT2LMHeadModel"
|
|
@@ -16,7 +17,10 @@
|
|
| 16 |
"n_inner": null,
|
| 17 |
"n_layer": 12,
|
| 18 |
"n_positions": 1024,
|
|
|
|
| 19 |
"resid_pdrop": 0.1,
|
|
|
|
|
|
|
| 20 |
"summary_activation": null,
|
| 21 |
"summary_first_dropout": 0.1,
|
| 22 |
"summary_proj_to_labels": true,
|
|
@@ -32,5 +36,8 @@
|
|
| 32 |
"top_p": 0.95
|
| 33 |
}
|
| 34 |
},
|
|
|
|
|
|
|
|
|
|
| 35 |
"vocab_size": 50257
|
| 36 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "gpt2",
|
| 3 |
"activation_function": "gelu_new",
|
| 4 |
"architectures": [
|
| 5 |
"GPT2LMHeadModel"
|
|
|
|
| 17 |
"n_inner": null,
|
| 18 |
"n_layer": 12,
|
| 19 |
"n_positions": 1024,
|
| 20 |
+
"reorder_and_upcast_attn": false,
|
| 21 |
"resid_pdrop": 0.1,
|
| 22 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 23 |
+
"scale_attn_weights": true,
|
| 24 |
"summary_activation": null,
|
| 25 |
"summary_first_dropout": 0.1,
|
| 26 |
"summary_proj_to_labels": true,
|
|
|
|
| 36 |
"top_p": 0.95
|
| 37 |
}
|
| 38 |
},
|
| 39 |
+
"torch_dtype": "float32",
|
| 40 |
+
"transformers_version": "4.16.2",
|
| 41 |
+
"use_cache": true,
|
| 42 |
"vocab_size": 50257
|
| 43 |
}
|
merges.txt
CHANGED
|
@@ -1,4 +1,4 @@
|
|
| 1 |
-
#version: 0.2
|
| 2 |
Ġ t
|
| 3 |
Ġ a
|
| 4 |
h e
|
|
|
|
| 1 |
+
#version: 0.2 - Trained by `huggingface/tokenizers`
|
| 2 |
Ġ t
|
| 3 |
Ġ a
|
| 4 |
h e
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:252aa2605584f92a5d64265a92c8ae7b84a76053bf48ec12e573ffb551cca675
|
| 3 |
+
size 510403817
|
special_tokens_map.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"bos_token":
|
|
|
|
| 1 |
+
{"bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "unk_token": "<|endoftext|>"}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"model_max_length": 1024}
|
|
|
|
| 1 |
+
{"unk_token": "<|endoftext|>", "bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "add_prefix_space": false, "model_max_length": 1024, "special_tokens_map_file": null, "name_or_path": "gpt2", "tokenizer_class": "GPT2Tokenizer"}
|
flax_model.msgpack → training_args.bin
RENAMED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4f9d60ed44eba7365d28f56b7b19ca4419042c9f4e9c8e4a2948802895be466d
|
| 3 |
+
size 3055
|
vocab.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|