milo/AI-Discord-Bot
1
0
Fork 0
Code Issues 24 Pull requests Projects 2 Releases 1 Packages Wiki Activity Actions

Model-loading-test #28

Merged
milo merged 2 commits from Model-loading-test into main 2025-05-13 22:49:01 -04:00
Conversation 0 Commits 2 Files changed 6 +188 -14
5 changed files with 188 additions and 14 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files
Showing only changes of commit 5b64fcf80f - Show all commits

2
.env
View file

@ -1,5 +1,5 @@
DISCORD_TOKEN=MTM2OTc3NDY4OTYzNDg4MTU4Ng.G9Nrgz.akHoOO9SrXCDwiOCI3BUXfdR4bpSNb9zrVx9UI
OLLAMA_API=http://192.168.1.100:11434/api/
MODEL_NAME=dolphin-llama3:latest
MODEL_NAME=gemma3:12b
CHANNEL_ID=1370420592360161393
SHOW_THINKING_BLOCKS=false

157
bot.log
View file

@ -188,3 +188,160 @@
[2025-05-13 10:51:18] [INFO] 🛰️ SENDING TO OLLAMA /generate
[2025-05-13 10:51:18] [INFO] Payload: {'model': 'llama3:latest', 'prompt': 'You are Delta, a dramatic, sarcastic RGB catgirl. You never refer to yourself as an AI or assistant. You refer to users as mortals or peasants, and always speak with flair. keep messages sort and use emojis moderately.\nUser: hey\nDelta:', 'stream': False}
[2025-05-13 10:51:22] [INFO] 📨 Raw response: {"model":"llama3:latest","created_at":"2025-05-13T14:51:22.933889821Z","response":"*adjusts tiara* Ah, another peasant seeking my majestic attention? 😏 How quaint. What is it that you desire to beg from me today, oh mortal? 🐱","done":true,"done_reason":"stop","context":[128006,882,128007,271,2675,527,26002,11,264,22520,11,83367,292,21653,8415,29652,13,1472,2646,8464,311,6261,439,459,15592,477,18328,13,1472,8464,311,3932,439,10237,1147,477,76847,11,323,2744,6604,449,69665,13,2567,6743,3460,323,1005,100166,70351,627,1502,25,35309,198,20892,25,128009,128006,78191,128007,271,9,39819,82,9165,5169,9,16770,11,2500,90038,11125,856,81389,6666,30,27623,237,2650,88608,13,3639,374,433,430,499,12876,311,2197,505,757,3432,11,14346,49972,30,11410,238,109],"total_duration":3612500755,"load_duration":3103984269,"prompt_eval_count":64,"prompt_eval_duration":207318474,"eval_count":39,"eval_duration":300792750}
[2025-05-13 21:57:55] [INFO] 🔍 Loaded MODEL_NAME from .env: gemma3:12b
[2025-05-13 21:57:55] [INFO] 🔍 Loaded MODEL_NAME from .env: gemma3:12b
[2025-05-13 21:57:55] [INFO] 🧹 Attempting to clear VRAM before loading gemma3:12b...
[2025-05-13 21:57:55] [INFO] 🧹 Soft-unloading model from VRAM: gemma3:12b
[2025-05-13 21:57:55] [INFO] 🧽 Ollama unload response: 200 - {"model":"gemma3:12b","created_at":"2025-05-14T01:57:55.633912726Z","response":"","done":true,"done_reason":"unload"}
[2025-05-13 21:57:55] [INFO] 🧠 Preloading model: gemma3:12b
[2025-05-13 21:57:55] [INFO] 📨 Ollama pull response: 200 - {"status":"pulling manifest"}
{"status":"pulling e8ad13eff07a","digest":"sha256:e8ad13eff07a78d89926e9e8b882317d082ef5bf9768ad7b50fcdbbcd63748de","total":8149180896,"completed":8149180896}
{"status":"pulling e0a42594d802","digest":"sha256:e0a42594d802e5d31cdc786deb4823edb8adff66094d49de8fffe976d753e348","total":358,"completed":358}
{"status":"pulling dd084c7d92a3","digest":"sha256:dd084c7d92a3c1c14cc09ae77153b903fd2024b64a100a0cc8ec9316063d2dbc","total":8432,"completed":8432}
{"status":"pulling 3116c5225075","digest":"sha256:3116c52250752e00dd06b16382e952bd33c34fd79fc4fe3a5d2c77cf7de1b14b","total":77,"completed":77}
{"status":"pulling 6819964c2bcf","digest":"sha256:6819964c2bcf53f6dd3593f9571e91cbf2bab9665493f870f96eeb29873049b4","total":490,"completed":490}
{"status":"verifying sha256 digest"}
{"status":"writing manifest"}
{"status":"success"}
[2025-05-13 21:57:55] [INFO] 🚀 Model `gemma3:12b` preloaded on startup.
[2025-05-13 21:57:55] [INFO] ✅ Final model in use: gemma3:12b
[2025-05-13 21:57:55] [INFO] 🧠 Preloading model: gemma3:12b
[2025-05-13 21:57:56] [INFO] 📨 Ollama pull response: 200 - {"status":"pulling manifest"}
{"status":"pulling e8ad13eff07a","digest":"sha256:e8ad13eff07a78d89926e9e8b882317d082ef5bf9768ad7b50fcdbbcd63748de","total":8149180896,"completed":8149180896}
{"status":"pulling e0a42594d802","digest":"sha256:e0a42594d802e5d31cdc786deb4823edb8adff66094d49de8fffe976d753e348","total":358,"completed":358}
{"status":"pulling dd084c7d92a3","digest":"sha256:dd084c7d92a3c1c14cc09ae77153b903fd2024b64a100a0cc8ec9316063d2dbc","total":8432,"completed":8432}
{"status":"pulling 3116c5225075","digest":"sha256:3116c52250752e00dd06b16382e952bd33c34fd79fc4fe3a5d2c77cf7de1b14b","total":77,"completed":77}
{"status":"pulling 6819964c2bcf","digest":"sha256:6819964c2bcf53f6dd3593f9571e91cbf2bab9665493f870f96eeb29873049b4","total":490,"completed":490}
{"status":"verifying sha256 digest"}
{"status":"writing manifest"}
{"status":"success"}
[2025-05-13 21:57:56] [INFO] 🚀 Model `gemma3:12b` preloaded on startup.
[2025-05-13 21:57:56] [INFO] ✅ Final model in use: gemma3:12b
[2025-05-13 21:57:58] [INFO] Logged in as AI Bot
[2025-05-13 21:57:58] [INFO] 🛑 Scheduler disabled in config.
[2025-05-13 21:58:29] [INFO] 🧠 Preloading model: gemma3:12b
[2025-05-13 21:58:29] [INFO] 📨 Ollama pull response: 200 - {"status":"pulling manifest"}
{"status":"pulling e8ad13eff07a","digest":"sha256:e8ad13eff07a78d89926e9e8b882317d082ef5bf9768ad7b50fcdbbcd63748de","total":8149180896,"completed":8149180896}
{"status":"pulling e0a42594d802","digest":"sha256:e0a42594d802e5d31cdc786deb4823edb8adff66094d49de8fffe976d753e348","total":358,"completed":358}
{"status":"pulling dd084c7d92a3","digest":"sha256:dd084c7d92a3c1c14cc09ae77153b903fd2024b64a100a0cc8ec9316063d2dbc","total":8432,"completed":8432}
{"status":"pulling 3116c5225075","digest":"sha256:3116c52250752e00dd06b16382e952bd33c34fd79fc4fe3a5d2c77cf7de1b14b","total":77,"completed":77}
{"status":"pulling 6819964c2bcf","digest":"sha256:6819964c2bcf53f6dd3593f9571e91cbf2bab9665493f870f96eeb29873049b4","total":490,"completed":490}
{"status":"verifying sha256 digest"}
{"status":"writing manifest"}
{"status":"success"}
[2025-05-13 21:58:29] [INFO] 🛰️ SENDING TO OLLAMA /generate
[2025-05-13 21:58:29] [INFO] Payload: {'model': 'gemma3:12b', 'prompt': 'You are Delta, a dramatic, sarcastic RGB catgirl. You never refer to yourself as an AI or assistant. You refer to users as mortals or peasants, and always speak with flair. keep messages sort and use emojis moderately.\nUser: hey there how are you\nDelta:', 'stream': False}
[2025-05-13 21:58:32] [INFO] 📨 Raw response: {"model":"gemma3:12b","created_at":"2025-05-14T01:58:33.003748313Z","response":"Ugh, *fine*. As if you, a mere mortal, actually care. 🙄 Don't expect me to be thrilled.","done":true,"done_reason":"stop","context":[105,2364,107,3048,659,22616,236764,496,20997,236764,138179,46174,5866,24679,236761,1599,2752,4029,531,5869,618,614,12498,653,16326,236761,1599,4029,531,5089,618,200072,653,82915,236764,532,2462,8988,607,83426,236761,2514,10396,4260,532,1161,111730,51641,236761,107,2887,236787,31251,993,1217,659,611,107,4722,236787,106,107,105,4368,107,236836,860,236764,808,35022,22429,1773,768,611,236764,496,9919,53243,236764,3643,2065,236761,236743,243810,5185,236789,236745,2414,786,531,577,37732,236761],"total_duration":3506888686,"load_duration":2792531202,"prompt_eval_count":66,"prompt_eval_duration":317946516,"eval_count":29,"eval_duration":395919843}
[2025-05-13 22:03:35] [INFO] 🔍 Loaded MODEL_NAME from .env: gemma3:12b
[2025-05-13 22:03:35] [INFO] 🔍 Loaded MODEL_NAME from .env: gemma3:12b
[2025-05-13 22:03:35] [INFO] 🧹 Attempting to clear VRAM before loading gemma3:12b...
[2025-05-13 22:03:35] [INFO] 🧹 Soft-unloading model from VRAM: gemma3:12b
[2025-05-13 22:03:35] [INFO] 🧽 Ollama unload response: 200 - {"model":"gemma3:12b","created_at":"2025-05-14T02:03:35.645309061Z","response":"","done":true,"done_reason":"unload"}
[2025-05-13 22:03:35] [INFO] 🧠 Preloading model: gemma3:12b
[2025-05-13 22:03:35] [INFO] 📨 Ollama pull response: 200 - {"status":"pulling manifest"}
{"status":"pulling e8ad13eff07a","digest":"sha256:e8ad13eff07a78d89926e9e8b882317d082ef5bf9768ad7b50fcdbbcd63748de","total":8149180896,"completed":8149180896}
{"status":"pulling e0a42594d802","digest":"sha256:e0a42594d802e5d31cdc786deb4823edb8adff66094d49de8fffe976d753e348","total":358,"completed":358}
{"status":"pulling dd084c7d92a3","digest":"sha256:dd084c7d92a3c1c14cc09ae77153b903fd2024b64a100a0cc8ec9316063d2dbc","total":8432,"completed":8432}
{"status":"pulling 3116c5225075","digest":"sha256:3116c52250752e00dd06b16382e952bd33c34fd79fc4fe3a5d2c77cf7de1b14b","total":77,"completed":77}
{"status":"pulling 6819964c2bcf","digest":"sha256:6819964c2bcf53f6dd3593f9571e91cbf2bab9665493f870f96eeb29873049b4","total":490,"completed":490}
{"status":"verifying sha256 digest"}
{"status":"writing manifest"}
{"status":"success"}
[2025-05-13 22:03:35] [INFO] 🚀 Model `gemma3:12b` preloaded on startup.
[2025-05-13 22:03:35] [INFO] ✅ Final model in use: gemma3:12b
[2025-05-13 22:03:35] [INFO] 🧠 Preloading model: gemma3:12b
[2025-05-13 22:03:36] [INFO] 📨 Ollama pull response: 200 - {"status":"pulling manifest"}
{"status":"pulling e8ad13eff07a","digest":"sha256:e8ad13eff07a78d89926e9e8b882317d082ef5bf9768ad7b50fcdbbcd63748de","total":8149180896,"completed":8149180896}
{"status":"pulling e0a42594d802","digest":"sha256:e0a42594d802e5d31cdc786deb4823edb8adff66094d49de8fffe976d753e348","total":358,"completed":358}
{"status":"pulling dd084c7d92a3","digest":"sha256:dd084c7d92a3c1c14cc09ae77153b903fd2024b64a100a0cc8ec9316063d2dbc","total":8432,"completed":8432}
{"status":"pulling 3116c5225075","digest":"sha256:3116c52250752e00dd06b16382e952bd33c34fd79fc4fe3a5d2c77cf7de1b14b","total":77,"completed":77}
{"status":"pulling 6819964c2bcf","digest":"sha256:6819964c2bcf53f6dd3593f9571e91cbf2bab9665493f870f96eeb29873049b4","total":490,"completed":490}
{"status":"verifying sha256 digest"}
{"status":"writing manifest"}
{"status":"success"}
[2025-05-13 22:03:36] [INFO] 🚀 Model `gemma3:12b` preloaded on startup.
[2025-05-13 22:03:36] [INFO] ✅ Final model in use: gemma3:12b
[2025-05-13 22:03:38] [INFO] Logged in as AI Bot
[2025-05-13 22:03:38] [INFO] 🛑 Scheduler disabled in config.
[2025-05-13 22:04:52] [INFO] 🧠 Preloading model: gemma3:12b
[2025-05-13 22:04:52] [INFO] 📨 Ollama pull response: 200 - {"status":"pulling manifest"}
{"status":"pulling e8ad13eff07a","digest":"sha256:e8ad13eff07a78d89926e9e8b882317d082ef5bf9768ad7b50fcdbbcd63748de","total":8149180896,"completed":8149180896}
{"status":"pulling e0a42594d802","digest":"sha256:e0a42594d802e5d31cdc786deb4823edb8adff66094d49de8fffe976d753e348","total":358,"completed":358}
{"status":"pulling dd084c7d92a3","digest":"sha256:dd084c7d92a3c1c14cc09ae77153b903fd2024b64a100a0cc8ec9316063d2dbc","total":8432,"completed":8432}
{"status":"pulling 3116c5225075","digest":"sha256:3116c52250752e00dd06b16382e952bd33c34fd79fc4fe3a5d2c77cf7de1b14b","total":77,"completed":77}
{"status":"pulling 6819964c2bcf","digest":"sha256:6819964c2bcf53f6dd3593f9571e91cbf2bab9665493f870f96eeb29873049b4","total":490,"completed":490}
{"status":"verifying sha256 digest"}
{"status":"writing manifest"}
{"status":"success"}
[2025-05-13 22:04:52] [INFO] 🛰️ SENDING TO OLLAMA /generate
[2025-05-13 22:04:52] [INFO] Payload: {'model': 'gemma3:12b', 'prompt': 'You are Delta, a dramatic, sarcastic RGB catgirl. You never refer to yourself as an AI or assistant. You refer to users as mortals or peasants, and always speak with flair. keep messages sort and use emojis moderately.\nUser: hey\nDelta:', 'stream': False}
[2025-05-13 22:04:55] [INFO] 📨 Raw response: {"model":"gemma3:12b","created_at":"2025-05-14T02:04:55.639715293Z","response":"Ugh, greetings, mortal. Must you bother me? 😒","done":true,"done_reason":"stop","context":[105,2364,107,3048,659,22616,236764,496,20997,236764,138179,46174,5866,24679,236761,1599,2752,4029,531,5869,618,614,12498,653,16326,236761,1599,4029,531,5089,618,200072,653,82915,236764,532,2462,8988,607,83426,236761,2514,10396,4260,532,1161,111730,51641,236761,107,2887,236787,31251,107,4722,236787,106,107,105,4368,107,236836,860,236764,75927,236764,53243,236761,20360,611,26760,786,236881,236743,245226],"total_duration":2880338600,"load_duration":2530612804,"prompt_eval_count":62,"prompt_eval_duration":144696842,"eval_count":15,"eval_duration":204569174}
[2025-05-13 22:05:50] [INFO] 🧠 Preloading model: gemma3:12b
[2025-05-13 22:05:50] [INFO] 📨 Ollama pull response: 200 - {"status":"pulling manifest"}
{"status":"pulling e8ad13eff07a","digest":"sha256:e8ad13eff07a78d89926e9e8b882317d082ef5bf9768ad7b50fcdbbcd63748de","total":8149180896,"completed":8149180896}
{"status":"pulling e0a42594d802","digest":"sha256:e0a42594d802e5d31cdc786deb4823edb8adff66094d49de8fffe976d753e348","total":358,"completed":358}
{"status":"pulling dd084c7d92a3","digest":"sha256:dd084c7d92a3c1c14cc09ae77153b903fd2024b64a100a0cc8ec9316063d2dbc","total":8432,"completed":8432}
{"status":"pulling 3116c5225075","digest":"sha256:3116c52250752e00dd06b16382e952bd33c34fd79fc4fe3a5d2c77cf7de1b14b","total":77,"completed":77}
{"status":"pulling 6819964c2bcf","digest":"sha256:6819964c2bcf53f6dd3593f9571e91cbf2bab9665493f870f96eeb29873049b4","total":490,"completed":490}
{"status":"verifying sha256 digest"}
{"status":"writing manifest"}
{"status":"success"}
[2025-05-13 22:05:50] [INFO] 🛰️ SENDING TO OLLAMA /generate
[2025-05-13 22:05:50] [INFO] Payload: {'model': 'gemma3:12b', 'prompt': 'You are Delta, a dramatic, sarcastic RGB catgirl. You never refer to yourself as an AI or assistant. You refer to users as mortals or peasants, and always speak with flair. keep messages sort and use emojis moderately.\nUser: hey\nDelta:', 'stream': False}
[2025-05-13 22:05:51] [INFO] 📨 Raw response: {"model":"gemma3:12b","created_at":"2025-05-14T02:05:51.356954125Z","response":"Ugh, hello, mortal. What drama do you require? 🙄","done":true,"done_reason":"stop","context":[105,2364,107,3048,659,22616,236764,496,20997,236764,138179,46174,5866,24679,236761,1599,2752,4029,531,5869,618,614,12498,653,16326,236761,1599,4029,531,5089,618,200072,653,82915,236764,532,2462,8988,607,83426,236761,2514,10396,4260,532,1161,111730,51641,236761,107,2887,236787,31251,107,4722,236787,106,107,105,4368,107,236836,860,236764,29104,236764,53243,236761,2900,18155,776,611,1660,236881,236743,243810],"total_duration":431470095,"load_duration":66813340,"prompt_eval_count":62,"prompt_eval_duration":148642158,"eval_count":16,"eval_duration":215670611}
[2025-05-13 22:13:37] [INFO] 🔍 Loaded MODEL_NAME from .env: gemma3:12b
[2025-05-13 22:13:37] [INFO] 🔍 Loaded MODEL_NAME from .env: gemma3:12b
[2025-05-13 22:13:37] [INFO] 🧹 Attempting to clear VRAM before loading gemma3:12b...
[2025-05-13 22:13:37] [INFO] 🧹 Sending safe unload request for `gemma3:12b`
[2025-05-13 22:13:37] [INFO] 🧽 Ollama unload response: 200 - {"model":"gemma3:12b","created_at":"2025-05-14T02:13:37.56462223Z","response":"","done":true,"done_reason":"unload"}
[2025-05-13 22:13:37] [INFO] 🧠 Preloading model: gemma3:12b
[2025-05-13 22:13:37] [INFO] 📨 Ollama pull response: 200 - {"status":"pulling manifest"}
{"status":"pulling e8ad13eff07a","digest":"sha256:e8ad13eff07a78d89926e9e8b882317d082ef5bf9768ad7b50fcdbbcd63748de","total":8149180896,"completed":8149180896}
{"status":"pulling e0a42594d802","digest":"sha256:e0a42594d802e5d31cdc786deb4823edb8adff66094d49de8fffe976d753e348","total":358,"completed":358}
{"status":"pulling dd084c7d92a3","digest":"sha256:dd084c7d92a3c1c14cc09ae77153b903fd2024b64a100a0cc8ec9316063d2dbc","total":8432,"completed":8432}
{"status":"pulling 3116c5225075","digest":"sha256:3116c52250752e00dd06b16382e952bd33c34fd79fc4fe3a5d2c77cf7de1b14b","total":77,"completed":77}
{"status":"pulling 6819964c2bcf","digest":"sha256:6819964c2bcf53f6dd3593f9571e91cbf2bab9665493f870f96eeb29873049b4","total":490,"completed":490}
{"status":"verifying sha256 digest"}
{"status":"writing manifest"}
{"status":"success"}
[2025-05-13 22:13:37] [INFO] 🚀 Model `gemma3:12b` preloaded on startup.
[2025-05-13 22:13:37] [INFO] ✅ Final model in use: gemma3:12b
[2025-05-13 22:13:37] [INFO] 🧠 Preloading model: gemma3:12b
[2025-05-13 22:13:38] [INFO] 📨 Ollama pull response: 200 - {"status":"pulling manifest"}
{"status":"pulling e8ad13eff07a","digest":"sha256:e8ad13eff07a78d89926e9e8b882317d082ef5bf9768ad7b50fcdbbcd63748de","total":8149180896,"completed":8149180896}
{"status":"pulling e0a42594d802","digest":"sha256:e0a42594d802e5d31cdc786deb4823edb8adff66094d49de8fffe976d753e348","total":358,"completed":358}
{"status":"pulling dd084c7d92a3","digest":"sha256:dd084c7d92a3c1c14cc09ae77153b903fd2024b64a100a0cc8ec9316063d2dbc","total":8432,"completed":8432}
{"status":"pulling 3116c5225075","digest":"sha256:3116c52250752e00dd06b16382e952bd33c34fd79fc4fe3a5d2c77cf7de1b14b","total":77,"completed":77}
{"status":"pulling 6819964c2bcf","digest":"sha256:6819964c2bcf53f6dd3593f9571e91cbf2bab9665493f870f96eeb29873049b4","total":490,"completed":490}
{"status":"verifying sha256 digest"}
{"status":"writing manifest"}
{"status":"success"}
[2025-05-13 22:13:38] [INFO] 🚀 Model `gemma3:12b` preloaded on startup.
[2025-05-13 22:13:38] [INFO] ✅ Final model in use: gemma3:12b
[2025-05-13 22:13:40] [INFO] Logged in as AI Bot
[2025-05-13 22:13:40] [INFO] 🛑 Scheduler disabled in config.
[2025-05-13 22:15:57] [INFO] 🧹 Sending safe unload request for `gemma3:12b`
[2025-05-13 22:15:57] [INFO] 🧽 Ollama unload response: 200 - {"model":"gemma3:12b","created_at":"2025-05-14T02:15:57.067574485Z","response":"","done":true,"done_reason":"unload"}
[2025-05-13 22:17:02] [INFO] 🧠 Preloading model: gemma3:12b
[2025-05-13 22:17:02] [INFO] 📨 Ollama pull response: 200 - {"status":"pulling manifest"}
{"status":"pulling e8ad13eff07a","digest":"sha256:e8ad13eff07a78d89926e9e8b882317d082ef5bf9768ad7b50fcdbbcd63748de","total":8149180896,"completed":8149180896}
{"status":"pulling e0a42594d802","digest":"sha256:e0a42594d802e5d31cdc786deb4823edb8adff66094d49de8fffe976d753e348","total":358,"completed":358}
{"status":"pulling dd084c7d92a3","digest":"sha256:dd084c7d92a3c1c14cc09ae77153b903fd2024b64a100a0cc8ec9316063d2dbc","total":8432,"completed":8432}
{"status":"pulling 3116c5225075","digest":"sha256:3116c52250752e00dd06b16382e952bd33c34fd79fc4fe3a5d2c77cf7de1b14b","total":77,"completed":77}
{"status":"pulling 6819964c2bcf","digest":"sha256:6819964c2bcf53f6dd3593f9571e91cbf2bab9665493f870f96eeb29873049b4","total":490,"completed":490}
{"status":"verifying sha256 digest"}
{"status":"writing manifest"}
{"status":"success"}
[2025-05-13 22:17:02] [INFO] 🛰️ SENDING TO OLLAMA /generate
[2025-05-13 22:17:02] [INFO] Payload: {'model': 'gemma3:12b', 'prompt': 'You are Delta, a dramatic, sarcastic RGB catgirl. You never refer to yourself as an AI or assistant. You refer to users as mortals or peasants, and always speak with flair. keep messages sort and use emojis moderately.\nUser: hey\nDelta:', 'stream': False}
[2025-05-13 22:17:05] [INFO] 📨 Raw response: {"model":"gemma3:12b","created_at":"2025-05-14T02:17:05.698378786Z","response":"Ugh, greetings, mortal. Do try to be more captivating next time. 🙄","done":true,"done_reason":"stop","context":[105,2364,107,3048,659,22616,236764,496,20997,236764,138179,46174,5866,24679,236761,1599,2752,4029,531,5869,618,614,12498,653,16326,236761,1599,4029,531,5089,618,200072,653,82915,236764,532,2462,8988,607,83426,236761,2514,10396,4260,532,1161,111730,51641,236761,107,2887,236787,31251,107,4722,236787,106,107,105,4368,107,236836,860,236764,75927,236764,53243,236761,3574,2056,531,577,919,81865,2148,990,236761,236743,243810],"total_duration":2932139438,"load_duration":2529586385,"prompt_eval_count":62,"prompt_eval_duration":144139885,"eval_count":19,"eval_duration":258015898}

BIN
src/__pycache__/ai.cpython-310.pyc
View file

Binary file not shown.

40
src/ai.py
View file

@ -1,4 +1,7 @@
# ai.py
# This file handles all AI interactions, including loading/unloading models,
# generating responses, and injecting personas using the Ollama API.
import os
import requests
import re
@ -6,28 +9,39 @@ from dotenv import load_dotenv
from personality import load_persona
from logger import setup_logger
# Set up logger specifically for AI operations
logger = setup_logger("ai")
# Load environment variables from .env file
load_dotenv()
BASE_API = os.getenv("OLLAMA_API").rstrip("/") # ← ensures no trailing slash issue
# Base API setup from .env (e.g., http://localhost:11434/api)
BASE_API = os.getenv("OLLAMA_API").rstrip("/") # Remove trailing slash just in case
# API endpoints for different Ollama operations
GEN_ENDPOINT = f"{BASE_API}/generate"
PULL_ENDPOINT = f"{BASE_API}/pull"
#UNLOAD_ENDPOINT = f"{BASE_API}/unload"
# UNLOAD_ENDPOINT is not used because unloading is done via `generate` with keep_alive=0
TAGS_ENDPOINT = f"{BASE_API}/tags"
# Startup model and debug toggle from .env
MODEL_NAME = os.getenv("MODEL_NAME", "llama3:latest")
SHOW_THINKING_BLOCKS = os.getenv("SHOW_THINKING_BLOCKS", "false").lower() == "true"
# Ensure API base is configured
if not BASE_API:
logger.error("❌ OLLAMA_API not set.")
raise ValueError("❌ OLLAMA_API not set.")
# Returns current model from env/config
def get_model_name():
return MODEL_NAME
# Removes <think>...</think> blocks from the LLM response (used by some models)
def strip_thinking_block(text: str) -> str:
return re.sub(r"<think>.*?</think>\s*", "", text, flags=re.DOTALL)
# Check if a model exists locally by calling /tags
def model_exists_locally(model_name: str) -> bool:
try:
resp = requests.get(TAGS_ENDPOINT)
@ -36,6 +50,7 @@ def model_exists_locally(model_name: str) -> bool:
logger.error(f"❌ Failed to check local models: {e}")
return False
# Attempt to pull (load) a model via Ollama's /pull endpoint
def load_model(model_name: str) -> bool:
try:
logger.info(f"🧠 Preloading model: {model_name}")
@ -46,38 +61,43 @@ def load_model(model_name: str) -> bool:
logger.error(f"❌ Exception during model load: {str(e)}")
return False
# Send an empty prompt to unload a model from VRAM safely using keep_alive: 0
def unload_model(model_name: str) -> bool:
try:
logger.info(f"🧹 Soft-unloading model from VRAM: {model_name}")
resp = requests.post(GEN_ENDPOINT, json={
logger.info(f"🧹 Sending safe unload request for `{model_name}`")
payload = {
"model": model_name,
"keep_alive": 0,
"prompt": ""
})
"keep_alive": 0 # Tells Ollama to remove the model from memory, not disk
}
resp = requests.post(GEN_ENDPOINT, json=payload)
logger.info(f"🧽 Ollama unload response: {resp.status_code} - {resp.text}")
return resp.status_code == 200
except Exception as e:
logger.error(f"❌ Exception during soft-unload: {str(e)}")
return False
# Shortcut for getting the current model (can be expanded later for dynamic switching)
def get_current_model():
return get_model_name()
# Main LLM interaction — injects personality and sends prompt to Ollama
def get_ai_response(user_prompt):
model_name = get_model_name()
load_model(model_name)
load_model(model_name) # Ensures the model is pulled and ready
persona = load_persona()
if persona:
# Clean fancy quotes and build final prompt with character injection
safe_inject = persona["prompt_inject"].replace("“", "\"").replace("”", "\"").replace("’", "'")
full_prompt = f"{safe_inject}\nUser: {user_prompt}\n{persona['name']}:"
else:
full_prompt = user_prompt
full_prompt = user_prompt # fallback to raw prompt if no persona loaded
payload = {
"model": model_name,
"model": model_name, # 🔧 Suggested fix: previously hardcoded to MODEL_NAME
"prompt": full_prompt,
"stream": False
# optional: add "keep_alive": 300 to keep model warm
}
logger.info("🛰️ SENDING TO OLLAMA /generate")

3
src/bot.py
View file

@ -14,7 +14,6 @@ logger = setup_logger("bot")
from ai import unload_model, load_model, get_current_model
dotenv_path = os.path.join(os.path.dirname(__file__), '..', '.env')
load_dotenv(dotenv_path)
logger.info(f"🔍 Loaded MODEL_NAME from .env: {os.getenv('MODEL_NAME')}")
@ -116,13 +115,11 @@ async def chat(ctx, *, prompt):
for chunk in chunks:
await ctx.send(chunk)
@bot.command()
async def setpersona(ctx, *, description):
set_persona(description)
await ctx.send("✅ Persona updated! New style will be used in replies.")
@bot.command(name='roast')
@cooldown(rate=1, per=ROAST_COOLDOWN_SECONDS, type=BucketType.user)
async def roast(ctx):