מחולל קטעים מהרצאות טד

.gitattributes

@@ -29,3 +29,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+model/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,8 +1,8 @@
 ---
-title: SillyHebrewTedTalkSnippetGenerator
-emoji: 💻
-colorFrom: pink
-colorTo: purple
+title: מחולל קטעים מהרצאות טד
+emoji: 🧑‍🏫
+colorFrom: green
+colorTo: pink
 sdk: gradio
 sdk_version: 3.1.7
 app_file: app.py

app.py

@@ -0,0 +1,45 @@
+import gradio as gr
+from transformers import pipeline
+
+title = "מחולל קטעים מהרצאות טד"
+description = ""
+article = "<p>Fine tuned <a href=\"https://huggingface.co/Norod78/hebrew-gpt_neo-small\">Norod78/hebrew-gpt_neo-small</a> upon a formatted <a href=\"https://www.kaggle.com/datasets/miguelcorraljr/ted-ultimate-dataset\"> TED – Ultimate Dataset</a> (Hebrew)</p>"
+
+model_id = "./model"
+text_generator = pipeline('text-generation', model=model_id, tokenizer=model_id)
+max_length = 128
+top_k = 40
+top_p = 0.92
+temperature = 1.0
+
+def text_generation(input_text = None):
+    if input_text == None or len(input_text) == 0:
+        input_text = "\t\""
+    else:
+        if input_text.startswith("<|startoftext|>") == False:            
+            input_text ="<|startoftext|>" + input_text
+    generated_text = text_generator(input_text,
+    max_length=max_length,
+    top_k=top_k, 
+    top_p=top_p,
+    temperature=temperature,
+    do_sample=True,
+    repetition_penalty=2.0,
+    bos_token='<|startoftext|>',
+    eos_token='<|endoftext|>', 
+    pad_token='<|pad|>',
+    unknown_token = '<|unknown|>',
+    num_return_sequences=1)
+    parsed_text = generated_text[0]["generated_text"].replace("<|startoftext|>", "").replace("\r","").replace("\n\n", "\n").replace("\t", " ").replace("<|pad|>", " * ").replace("\"\"", "\"")
+    return parsed_text
+gr.Interface(
+    text_generation,    
+    inputs=gr.Textbox(lines=1, label="הזינו פה ציטוט פתיחה, או השאירו ריק. מה שבא לכם", elem_id="input_text"),
+    outputs=gr.Textbox(type="auto", label="פה מופיע הטקסט שהמחולל יוצר", elem_id="output_text"),
+    css="#output_text{direction: rtl} #input_text{direction: rtl}",
+    title=title,
+    description=description,
+    article=article,
+    theme="default",
+    allow_flagging=False,
+).launch()

model/added_tokens.json

@@ -0,0 +1,6 @@
+{
+  "<|endoftext|>": 50257,
+  "<|pad|>": 50260,
+  "<|startoftext|>": 50258,
+  "<|unknown|>": 50259
+}

model/config.json

@@ -0,0 +1,54 @@
+{
+  "_name_or_path": "Norod78/hebrew-gpt_neo-small",
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPTNeoForCausalLM"
+  ],
+  "attention_dropout": 0,
+  "attention_layers": [
+    "global",
+    "global",
+    "global",
+    "global",
+    "global",
+    "global",
+    "global",
+    "global",
+    "global",
+    "global",
+    "global",
+    "global"
+  ],
+  "attention_types": [
+    [
+      [
+        "global"
+      ],
+      12
+    ]
+  ],
+  "bos_token_id": 50256,
+  "embed_dropout": 0,
+  "eos_token_id": 50256,
+  "gradient_checkpointing": false,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": null,
+  "layer_norm_epsilon": 1e-05,
+  "max_position_embeddings": 2048,
+  "model_type": "gpt_neo",
+  "num_heads": 12,
+  "num_layers": 12,
+  "pad_token_id": 50256,
+  "resid_dropout": 0,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.20.1",
+  "use_cache": true,
+  "vocab_size": 50261,
+  "window_size": 256
+}

model/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d44ea32f47ad8ed6713087424ff7f165616f6c2c7f5b35e3b1a872a867ed6a6
+size 551197393

model/special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "<|startoftext|>",
+  "eos_token": "<|endoftext|>",
+  "pad_token": "<|pad|>",
+  "unk_token": "<unk>"
+}

model/tokenizer_config.json

@@ -0,0 +1,23 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "bos_token": "<|startoftext|>",
+  "do_lower_case": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "full_tokenizer_file": null,
+  "max_len": 1024,
+  "name_or_path": "Norod78/hebrew-gpt_neo-small",
+  "pad_token": "<|pad|>",
+  "special_tokens_map_file": "special_tokens_map.json",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": {
+    "__type": "AddedToken",
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unknown_token": "<|unknown|>"
+}

requirements.txt

@@ -0,0 +1,5 @@
+gradio
+torch
+transformers
+tokenizers
+