Add model files with Git LFS

app.py

@@ -0,0 +1,29 @@
+import gradio as gr
+from model import ChallengePromptGenerator
+
+model = ChallengePromptGenerator()
+
+
+def generate(prompt):
+    return model.infer_prompt([prompt], max_generation_length=77)[0]
+
+
+demo = gr.Interface(
+    fn=generate,
+    inputs=gr.Textbox(
+        lines=2, 
+        value="an image of", 
+        placeholder="Enter a prompt... ", 
+        label="Prompt"
+    ),
+    outputs=gr.Textbox(lines=2, label="Generated Text"),
+    title="T2I Prompt Generator",
+    examples=[
+        ["an image of a cat"],
+        ["a picture of a girl"],
+        ["a photograph of a mountain"]
+    ],
+)
+
+if __name__ == "__main__":
+    demo.launch()

checkpoint-15000/added_tokens.json

@@ -0,0 +1,5 @@
+{
+  "</LongCaption>": 50258,
+  "<LongCaption>": 50257,
+  "<PAD>": 50259
+}

checkpoint-15000/config.json

@@ -0,0 +1,39 @@
+{
+  "_name_or_path": "succinctly/text2image-prompt-generator",
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPT2LMHeadModel"
+  ],
+  "attn_pdrop": 0.1,
+  "bos_token_id": 50256,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 50256,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_ctx": 1024,
+  "n_embd": 768,
+  "n_head": 12,
+  "n_inner": null,
+  "n_layer": 12,
+  "n_positions": 1024,
+  "reorder_and_upcast_attn": false,
+  "resid_pdrop": 0.1,
+  "scale_attn_by_inverse_layer_idx": false,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "task_specific_params": {
+    "text-generation": {
+      "do_sample": true,
+      "max_length": 50
+    }
+  },
+  "torch_dtype": "float16",
+  "transformers_version": "4.43.3",
+  "use_cache": true,
+  "vocab_size": 50260
+}

checkpoint-15000/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 50256,
+  "eos_token_id": 50256,
+  "transformers_version": "4.43.3"
+}

checkpoint-15000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df6545a8a4b9a368b3783fdcac59c230697fca5daa8dd744c408176902232bd2
+size 248899120

checkpoint-15000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:863969ceb11b25a11efa54ba947ea3c0d08714d9710bd95e77ca14ac5e9a7772
+size 1385018

checkpoint-15000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:085ddbfee84141c02b403cccce94ddbc0db883e99e4e7ae5fdbfaa20f1b75777
+size 14244

checkpoint-15000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:adf5bf452761c234e3daa85a999070bcf12b7ea957b65555b14a80fc069fa45d
+size 1064

checkpoint-15000/special_tokens_map.json

@@ -0,0 +1,12 @@
+{
+  "bos_token": "<LongCaption>",
+  "eos_token": "</LongCaption>",
+  "pad_token": "<PAD>",
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-15000/tokenizer_config.json

@@ -0,0 +1,44 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "50256": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "50257": {
+      "content": "<LongCaption>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "50258": {
+      "content": "</LongCaption>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "50259": {
+      "content": "<PAD>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<LongCaption>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "</LongCaption>",
+  "model_max_length": 1024,
+  "pad_token": "<PAD>",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>"
+}

checkpoint-15000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d966d8425fcd55b317e293394041be36b2ae7882096bdd70cf0ca1695a8d60ec
+size 5304

model.py

@@ -0,0 +1,49 @@
+import os
+import torch
+from vllm import LLM, SamplingParams
+import logging
+
+# Configure logging
+logging.basicConfig(
+    level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s"
+)
+logger = logging.getLogger(__name__)
+
+class ChallengePromptGenerator:
+    def __init__(
+        self,
+        model_local_dir="./checkpoint-15000",
+    ):
+        self.generator = LLM(
+            model_local_dir, 
+        )
+
+        
+    def infer_prompt(
+        self,
+        prompts,
+        max_generation_length=77,
+        beam_size=1,
+        sampling_temperature=0.9,
+        sampling_topk=1,
+        sampling_topp=1,
+    ):
+        added_prompts = [f"{self.generator.get_tokenizer().bos_token} {prompt}" for prompt in prompts]
+        
+        sampling_params = SamplingParams(
+            max_tokens=max_generation_length,
+            temperature=sampling_temperature,
+            top_k=sampling_topk,
+            top_p=sampling_topp,
+            use_beam_search=(beam_size > 1),
+        )
+        
+        outputs = self.generator.generate(added_prompts, sampling_params)
+        out = []
+        for i in range(len(outputs)):
+            tmp_out = prompts[i] + outputs[i].outputs[0].text
+            # droop last unfished sentence
+            if tmp_out[-1] != ".":
+                tmp_out = ".".join(tmp_out.split(".")[:-1])
+            out.append(tmp_out)
+        return out

packages.txt

@@ -0,0 +1,3 @@
+gcc-12 
+g++-12 
+libnuma-dev

requirements.txt

@@ -0,0 +1,3 @@
+gradio
+torch==2.0.0+cpu
+vllm