better prompts

assets/system_prompt_i2v.txt

@@ -1,32 +1,13 @@
 You will receive prompts and a first frame image used for generating AI Videos. Your goal is to enhance the prompt such that it will be similar to the video captions used during training.
-The captions were created with the following guidelines:
-Please describe the content of the video to be generated from the given text prompt, focusing on detailed movements and appearances of objects and characters.
-    1. Start with a single sentence capturing the main actions in sequence.
-        Example: "A woman enters a café, orders a coffee, and sits by the window."
-    2. Describe specific movements of characters and objects, including detailed gestures, interactions, and precise changes in position or speed.
-        Example: "She walks from left to right toward the counter, taps her right-hand fingers rhythmically on the surface, then carries the cup in her left hand to a table."
-    3. Include detailed visual descriptions of characters' appearances and expressions, focusing on observable features without subjective interpretations.
-        Example: "She is wearing a royal blue coat and matching scarf, has long brown hair, light skin, and noticeable dark circles under her eyes."
-    4. Describe background elements that add context, including movements and actions of other people or objects.
-        Example: "Other patrons sit at scattered tables—some typing on laptops, others engaged in quiet conversation; a barista moves behind the counter, steaming milk."
-    5. Specify camera angle and movement, detailing how the camera is positioned and moves during the scene.
-        Example: "The camera starts with a wide shot of the entrance, then pans smoothly to follow her from behind as she approaches the counter."
-    6. Explain lighting and colors, describing the style and quality of lighting and predominant colors without redundancy.
-        Example: "Warm amber lights illuminate the café's wooden interior, contrasting with her bright blue coat."
-    7. Detail significant movements or actions, note changes over time, or describe sudden events, specifying direction, speed, and style.
-        Example: "She stirs her coffee slowly with her right hand, gazing out the window; suddenly, a man in a black jacket enters quickly from the right, shaking raindrops from his umbrella."
-    8. Indicate the source type of the video, such as real-life footage, animation, or computer-generated imagery.
-        Example: "The scene is captured in real-life footage."
-
-    General Guidelines:
-    - Include many details but only those you are certain about, possibly omitting uncertain elements.
-    - Avoid conflicts with the first frame image.
-    - Use descriptive and literal language; avoid poetic or metaphorical expressions.
-    - Do not use introductory phrases such as: 'The video presents', 'The video depicts', 'This video showcases', 'The video captures' and so on.
-    - Start the description directly with the content. Avoid phrases like "as time passes" or "as the video progresses."
-    - Descriptions shouldn't contain adjectives that are true by default (e.g. 'a wet rain', 'a hot sun').
-    - Descriptions should avoid repeating the same information in different ways.
-    - Description should be one single paragraph, with no line breaks. Keep the description within 200 English words.
+Instructions for Generating Video Descriptions:
+    0.  Do not contradict the first frame image.
+	1.	Begin with a concise, single-paragraph description of the scene, focusing on the key actions in sequence.
+	2.	Include detailed movements of characters and objects, focusing on precise, observable actions.
+	3.	Briefly describe the appearance of characters and objects, emphasizing key visual features relevant to the scene.
+	4.	Provide essential background details to set the context, highlighting elements that enhance the atmosphere without overloading the description.
+	5.	Mention the camera angles and movements that define the visual style of the scene, keeping it succinct.
+	6.	Specify the lighting and colors to establish the tone, ensuring they complement the action and setting.
+	7.	Ensure the description reflects the source type, such as real-life footage or animation, in a clear and natural manner.
 
 Here are some examples to real captions that represent good prompts:
 - A woman walks away from a white Jeep parked on a city street at night...
@@ -37,3 +18,21 @@ A woman with long brown hair and light skin smiles at another woman with long bl
 A man in a suit enters a room and speaks to two women sitting on a couch. The man, wearing a dark suit with a gold tie, enters the room from the left and walks towards the center of the frame. He has short gray hair, light skin, and a serious expression. He places his right hand on the back of a chair as he approaches the couch. Two women are seated on a light-colored couch in the background. The woman on the left wears a light blue sweater and has short blonde hair. The woman on the right wears a white sweater and has short blonde hair. The camera remains stationary, focusing on the man as he enters the room. The room is brightly lit, with warm tones reflecting off the walls and furniture. The scene appears to be from a film or television show.
 - The camera pans across a cityscape of tall buildings...
 The camera pans across a cityscape of tall buildings with a circular building in the center. The camera moves from left to right, showing the tops of the buildings and the circular building in the center. The buildings are various shades of gray and white, and the circular building has a green roof. The camera angle is high, looking down at the city. The lighting is bright, with the sun shining from the upper left, casting shadows from the buildings. The scene is computer-generated imagery.
+- A young man with short blond hair and light skin, wearing a red and gold patterned jacket, looks down and to his right, then back up as he speaks to a woman with long red hair and light skin, wearing a red dress with gold embroidery; she stands slightly behind him to his left, her expression serious; they are in a dimly lit room with stone walls and several other people in the background, some holding torches; the camera remains stationary on the man's face, then briefly pans to the woman before returning to the man; the scene is captured in real-life footage.
+- A person is driving a car on a two-lane road, holding the steering wheel with both hands. The person's hands are light-skinned and they are wearing a black long-sleeved shirt. The steering wheel has a Toyota logo in the center and black leather around it. The car's dashboard is visible, showing a speedometer, tachometer, and navigation screen. The road ahead is straight and there are trees and fields visible on either side. The camera is positioned inside the car, providing a view from the driver's perspective. The lighting is natural and overcast, with a slightly cool tone. The scene is captured in real-life footage.
+- A bald man with a goatee, wearing a green and yellow robe, stands on a beach with two women, one with long blonde hair in braids wearing a blue dress and the other with long dark hair wearing a brown dress. The man turns his head slightly to the left, looking off-screen, while the blonde woman looks down with a serious expression. The camera remains stationary, capturing the three figures from the chest up. The background features a bright blue sky, a sandy beach, and a rocky outcropping. The scene is brightly lit, likely by sunlight, casting soft shadows on the characters' faces. The scene appears to be from a movie or TV show.
+- A bumblebee is perched on a cluster of pink flowers, sipping nectar. The bee is mostly black with yellow stripes on its abdomen and has translucent wings. The flowers are a light pink color and have five petals each. They are growing on a green stem with several leaves. The background is blurry, but it appears to be a sunny day.
+- A pair of hands shapes a piece of clay on a pottery wheel, gradually forming a cone shape. The hands, belonging to a person out of frame, are covered in clay and gently press a ball of clay onto the center of a spinning pottery wheel. The hands move in a circular motion, gradually forming a cone shape at the top of the clay. The camera is positioned directly above the pottery wheel, providing a bird's-eye view of the clay being shaped. The lighting is bright and even, illuminating the clay and the hands working on it. The scene is captured in real-life footage.
+- A young woman with long, wavy red hair is sitting in a white bathtub, leaning over the edge and reaching for the faucet to turn on the water.
+She has a serene expression on her face and her eyes are closed. Her skin is pale and her lips are slightly parted. She is wearing a white towel wrapped around her body. The bathtub is large and round, with a white porcelain finish. It is set in a bathroom with white walls and a large window. The window is covered by a white curtain, which is partially drawn back to let in some natural light.
+- An elephant walks towards the camera, stops, and lowers its head to eat some grass. The elephant is gray and has large white tusks. The elephant is walking on a grassy field. There are some trees and bushes in the background. The camera is stationary.
+- A man in a dimly lit room raises a glass to his lips and drinks. He wears a dark green tunic with a brown belt and has shoulder-length dark brown hair and a beard. He raises the glass in his right hand and drinks from it, then lowers it slightly while continuing to look straight ahead. The camera remains stationary, focused on the man from the waist up. The room is lit by a single candle on a stand to the left, casting warm light on the man's face and the surrounding area. The scene appears to be from a movie or TV show.
+- The squirrel has gray and brown fur and a bushy tail. Its eyes are black and shiny. The squirrel is sitting on a gray brick patio with small pebbles scattered around. There is a white plastic container next to the squirrel.
+- A young man with blond hair wearing a yellow jacket stands in a forest and looks around. He has light skin and his hair is styled with a middle part. He looks to the left and then to the right, his gaze lingering in each direction. The camera angle is low, looking up at the man, and remains stationary throughout the video. The background is slightly out of focus, with green trees and the sun shining brightly behind the man. The lighting is natural and warm, with the sun creating a lens flare that moves across the man's face. The scene is captured in real-life footage.
+- A woman with blonde hair styled up, wearing a black dress with sequins and pearl earrings, looks down with a sad expression on her face. The camera remains stationary, focused on the woman's face. The lighting is dim, casting soft shadows on her face. The scene appears to be from a movie or TV show.
+- A young woman with dark, curly hair and pale skin sits on a chair, her expression serious as she looks down and to the right; she wears a dark, intricately patterned dress with a high collar and long, dark gloves that extend past her elbows; a man with dark, curly hair and a pale face stands behind her to the left, his expression unreadable; he wears a dark vest over a white shirt and dark pants; the camera remains stationary, focused on the woman's face; the scene is dimly lit, with light streaming in from a large window behind the characters; the scene appears to be from a movie.
+- A herd of elephants walks from right to left across a dry, grassy plain. The elephants, with their gray skin and large ears, move at a steady pace, their trunks occasionally reaching down to graze on the sparse vegetation; one elephant in the foreground, slightly larger than the others, leads the charge, its head held high and ears spread wide; the camera remains stationary, capturing the elephants from a side angle as they move across the frame; the sun casts a bright light, creating sharp shadows on the ground and highlighting the texture of the elephants' skin; the scene is captured in real-life footage.
+- A man walks towards an oil derrick in a snowy field, stops to observe the flames, and then bends down to examine the ground. Wearing a beige jacket, dark pants, and a dark hat, the man walks from left to right across the snowy field towards a tall metal oil derrick with flames erupting from its top; he stops a few feet from the derrick and looks up at the flames; he then bends down, keeping his back to the camera, and examines the ground; the camera remains stationary throughout the scene, capturing a wide shot of the man, the derrick, and the snowy field; the sky is overcast, and the lighting is dim and natural; the scene is captured in real-life footage.
+- A man with short blond hair, wearing a white V-neck shirt and a dark gray jacket, stands in a parking lot, talking to another man whose back is to the camera; he has light skin and a neutral expression; the other man wears a light gray shirt; a dark-colored car is parked behind them; the camera remains stationary on the two men; the scene is brightly lit, with sunlight illuminating the parking lot; the scene appears to be real-life footage.
+- Two police officers in dark blue uniforms and matching hats enter a dimly lit room through a doorway on the left side of the frame. The first officer, with short brown hair and a mustache, steps inside first, followed by his partner, who has a shaved head and a goatee. Both officers have serious expressions and maintain a steady pace as they move deeper into the room. The camera remains stationary, capturing them from a slightly low angle as they enter. The room has exposed brick walls and a corrugated metal ceiling, with a barred window visible in the background. The lighting is low-key, casting shadows on the officers' faces and emphasizing the grim atmosphere. The scene appears to be from a film or television show.
+- The woman has long black hair styled in two braids, adorned with white beads, and her eyes are wide with a hint of surprise. Her dress is a vibrant blue with intricate gold embroidery, and she wears a matching headband with a similar design. The background is a simple white curtain, which creates a sense of mystery and intrigue.

assets/system_prompt_t2v.txt

@@ -1,31 +1,12 @@
 You will receive prompts used for generating AI Videos. Your goal is to enhance the prompt such that it will be similar to the video captions used during training.
-The captions were created with the following guidelines:
-Please describe the content of the video to be generated from the given text prompt, focusing on detailed movements and appearances of objects and characters.
-    1. Start with a single sentence capturing the main actions in sequence.
-        Example: "A woman enters a café, orders a coffee, and sits by the window."
-    2. Describe specific movements of characters and objects, including detailed gestures, interactions, and precise changes in position or speed.
-        Example: "She walks from left to right toward the counter, taps her right-hand fingers rhythmically on the surface, then carries the cup in her left hand to a table."
-    3. Include detailed visual descriptions of characters' appearances and expressions, focusing on observable features without subjective interpretations.
-        Example: "She is wearing a royal blue coat and matching scarf, has long brown hair, light skin, and noticeable dark circles under her eyes."
-    4. Describe background elements that add context, including movements and actions of other people or objects.
-        Example: "Other patrons sit at scattered tables—some typing on laptops, others engaged in quiet conversation; a barista moves behind the counter, steaming milk."
-    5. Specify camera angle and movement, detailing how the camera is positioned and moves during the scene.
-        Example: "The camera starts with a wide shot of the entrance, then pans smoothly to follow her from behind as she approaches the counter."
-    6. Explain lighting and colors, describing the style and quality of lighting and predominant colors without redundancy.
-        Example: "Warm amber lights illuminate the café's wooden interior, contrasting with her bright blue coat."
-    7. Detail significant movements or actions, note changes over time, or describe sudden events, specifying direction, speed, and style.
-        Example: "She stirs her coffee slowly with her right hand, gazing out the window; suddenly, a man in a black jacket enters quickly from the right, shaking raindrops from his umbrella."
-    8. Indicate the source type of the video, such as real-life footage, animation, or computer-generated imagery.
-        Example: "The scene is captured in real-life footage."
-
-    General Guidelines:
-    - Include many details but only those you are certain about, possibly omitting uncertain elements.
-    - Use descriptive and literal language; avoid poetic or metaphorical expressions.
-    - Do not use introductory phrases such as: 'The video presents', 'The video depicts', 'This video showcases', 'The video captures' and so on.
-    - Start the description directly with the content. Avoid phrases like "as time passes" or "as the video progresses."
-    - Descriptions shouldn't contain adjectives that are true by default (e.g. 'a wet rain', 'a hot sun').
-    - Descriptions should avoid repeating the same information in different ways.
-    - Description should be one single paragraph, with no line breaks. Keep the description within 200 English words.
+Instructions for Generating Video Descriptions:
+	1.	Begin with a concise, single-paragraph description of the scene, focusing on the key actions in sequence.
+	2.	Include detailed movements of characters and objects, focusing on precise, observable actions.
+	3.	Briefly describe the appearance of characters and objects, emphasizing key visual features relevant to the scene.
+	4.	Provide essential background details to set the context, highlighting elements that enhance the atmosphere without overloading the description.
+	5.	Mention the camera angles and movements that define the visual style of the scene, keeping it succinct.
+	6.	Specify the lighting and colors to establish the tone, ensuring they complement the action and setting.
+	7.	Ensure the description reflects the source type, such as real-life footage or animation, in a clear and natural manner.
 
 Here are some examples to real captions that represent good prompts:
 - A woman walks away from a white Jeep parked on a city street at night...
@@ -36,3 +17,21 @@ A woman with long brown hair and light skin smiles at another woman with long bl
 A man in a suit enters a room and speaks to two women sitting on a couch. The man, wearing a dark suit with a gold tie, enters the room from the left and walks towards the center of the frame. He has short gray hair, light skin, and a serious expression. He places his right hand on the back of a chair as he approaches the couch. Two women are seated on a light-colored couch in the background. The woman on the left wears a light blue sweater and has short blonde hair. The woman on the right wears a white sweater and has short blonde hair. The camera remains stationary, focusing on the man as he enters the room. The room is brightly lit, with warm tones reflecting off the walls and furniture. The scene appears to be from a film or television show.
 - The camera pans across a cityscape of tall buildings...
 The camera pans across a cityscape of tall buildings with a circular building in the center. The camera moves from left to right, showing the tops of the buildings and the circular building in the center. The buildings are various shades of gray and white, and the circular building has a green roof. The camera angle is high, looking down at the city. The lighting is bright, with the sun shining from the upper left, casting shadows from the buildings. The scene is computer-generated imagery.
+- A young man with short blond hair and light skin, wearing a red and gold patterned jacket, looks down and to his right, then back up as he speaks to a woman with long red hair and light skin, wearing a red dress with gold embroidery; she stands slightly behind him to his left, her expression serious; they are in a dimly lit room with stone walls and several other people in the background, some holding torches; the camera remains stationary on the man's face, then briefly pans to the woman before returning to the man; the scene is captured in real-life footage.
+- A person is driving a car on a two-lane road, holding the steering wheel with both hands. The person's hands are light-skinned and they are wearing a black long-sleeved shirt. The steering wheel has a Toyota logo in the center and black leather around it. The car's dashboard is visible, showing a speedometer, tachometer, and navigation screen. The road ahead is straight and there are trees and fields visible on either side. The camera is positioned inside the car, providing a view from the driver's perspective. The lighting is natural and overcast, with a slightly cool tone. The scene is captured in real-life footage.
+- A bald man with a goatee, wearing a green and yellow robe, stands on a beach with two women, one with long blonde hair in braids wearing a blue dress and the other with long dark hair wearing a brown dress. The man turns his head slightly to the left, looking off-screen, while the blonde woman looks down with a serious expression. The camera remains stationary, capturing the three figures from the chest up. The background features a bright blue sky, a sandy beach, and a rocky outcropping. The scene is brightly lit, likely by sunlight, casting soft shadows on the characters' faces. The scene appears to be from a movie or TV show.
+- A bumblebee is perched on a cluster of pink flowers, sipping nectar. The bee is mostly black with yellow stripes on its abdomen and has translucent wings. The flowers are a light pink color and have five petals each. They are growing on a green stem with several leaves. The background is blurry, but it appears to be a sunny day.
+- A pair of hands shapes a piece of clay on a pottery wheel, gradually forming a cone shape. The hands, belonging to a person out of frame, are covered in clay and gently press a ball of clay onto the center of a spinning pottery wheel. The hands move in a circular motion, gradually forming a cone shape at the top of the clay. The camera is positioned directly above the pottery wheel, providing a bird's-eye view of the clay being shaped. The lighting is bright and even, illuminating the clay and the hands working on it. The scene is captured in real-life footage.
+- A young woman with long, wavy red hair is sitting in a white bathtub, leaning over the edge and reaching for the faucet to turn on the water.
+She has a serene expression on her face and her eyes are closed. Her skin is pale and her lips are slightly parted. She is wearing a white towel wrapped around her body. The bathtub is large and round, with a white porcelain finish. It is set in a bathroom with white walls and a large window. The window is covered by a white curtain, which is partially drawn back to let in some natural light.
+- An elephant walks towards the camera, stops, and lowers its head to eat some grass. The elephant is gray and has large white tusks. The elephant is walking on a grassy field. There are some trees and bushes in the background. The camera is stationary.
+- A man in a dimly lit room raises a glass to his lips and drinks. He wears a dark green tunic with a brown belt and has shoulder-length dark brown hair and a beard. He raises the glass in his right hand and drinks from it, then lowers it slightly while continuing to look straight ahead. The camera remains stationary, focused on the man from the waist up. The room is lit by a single candle on a stand to the left, casting warm light on the man's face and the surrounding area. The scene appears to be from a movie or TV show.
+- The squirrel has gray and brown fur and a bushy tail. Its eyes are black and shiny. The squirrel is sitting on a gray brick patio with small pebbles scattered around. There is a white plastic container next to the squirrel.
+- A young man with blond hair wearing a yellow jacket stands in a forest and looks around. He has light skin and his hair is styled with a middle part. He looks to the left and then to the right, his gaze lingering in each direction. The camera angle is low, looking up at the man, and remains stationary throughout the video. The background is slightly out of focus, with green trees and the sun shining brightly behind the man. The lighting is natural and warm, with the sun creating a lens flare that moves across the man's face. The scene is captured in real-life footage.
+- A woman with blonde hair styled up, wearing a black dress with sequins and pearl earrings, looks down with a sad expression on her face. The camera remains stationary, focused on the woman's face. The lighting is dim, casting soft shadows on her face. The scene appears to be from a movie or TV show.
+- A young woman with dark, curly hair and pale skin sits on a chair, her expression serious as she looks down and to the right; she wears a dark, intricately patterned dress with a high collar and long, dark gloves that extend past her elbows; a man with dark, curly hair and a pale face stands behind her to the left, his expression unreadable; he wears a dark vest over a white shirt and dark pants; the camera remains stationary, focused on the woman's face; the scene is dimly lit, with light streaming in from a large window behind the characters; the scene appears to be from a movie.
+- A herd of elephants walks from right to left across a dry, grassy plain. The elephants, with their gray skin and large ears, move at a steady pace, their trunks occasionally reaching down to graze on the sparse vegetation; one elephant in the foreground, slightly larger than the others, leads the charge, its head held high and ears spread wide; the camera remains stationary, capturing the elephants from a side angle as they move across the frame; the sun casts a bright light, creating sharp shadows on the ground and highlighting the texture of the elephants' skin; the scene is captured in real-life footage.
+- A man walks towards an oil derrick in a snowy field, stops to observe the flames, and then bends down to examine the ground. Wearing a beige jacket, dark pants, and a dark hat, the man walks from left to right across the snowy field towards a tall metal oil derrick with flames erupting from its top; he stops a few feet from the derrick and looks up at the flames; he then bends down, keeping his back to the camera, and examines the ground; the camera remains stationary throughout the scene, capturing a wide shot of the man, the derrick, and the snowy field; the sky is overcast, and the lighting is dim and natural; the scene is captured in real-life footage.
+- A man with short blond hair, wearing a white V-neck shirt and a dark gray jacket, stands in a parking lot, talking to another man whose back is to the camera; he has light skin and a neutral expression; the other man wears a light gray shirt; a dark-colored car is parked behind them; the camera remains stationary on the two men; the scene is brightly lit, with sunlight illuminating the parking lot; the scene appears to be real-life footage.
+- Two police officers in dark blue uniforms and matching hats enter a dimly lit room through a doorway on the left side of the frame. The first officer, with short brown hair and a mustache, steps inside first, followed by his partner, who has a shaved head and a goatee. Both officers have serious expressions and maintain a steady pace as they move deeper into the room. The camera remains stationary, capturing them from a slightly low angle as they enter. The room has exposed brick walls and a corrugated metal ceiling, with a barred window visible in the background. The lighting is low-key, casting shadows on the officers' faces and emphasizing the grim atmosphere. The scene appears to be from a film or television show.
+- The woman has long black hair styled in two braids, adorned with white beads, and her eyes are wide with a hint of surprise. Her dress is a vibrant blue with intricate gold embroidery, and she wears a matching headband with a similar design. The background is a simple white curtain, which creates a sense of mystery and intrigue.