← العودة
Blog Post
عنوان المحادثة: MedGemma is a collection of Gemma 3 variants that are trained for perf...
التاريخ: 02.05.2026 - 03.05.2026
التصنيف: 🧠 الذكاء الاصطناعي وتعلم الآلة
إجمالي الرسائل: 46 | ياسر: 38 | M: 8
Yasser
MedGemma is a collection of Gemma 3 variants that are trained for performance on medical text and image comprehension.
Yasser
المسار ده فيه الملفات كاملة وبما فيها ملفات الإيفالويت/root/yasser/assignment1/assessment2/Assessment 2.0 Paid Flooring Challenge - TAKEOFF-52 Lovesacإنسخ الملفات للمكان الصحيح في ملفات العميل علشان أقدر أعمل عليها إيفالويت
المحادثة الكاملة - 02.05.2026 - 03.05.2026
ياسر
MedGemma is a collection of Gemma 3 variants that are trained for performance on medical text and image comprehension.
02.05.2026 11:25
ياسر
المسار ده فيه الملفات كاملة وبما فيها ملفات الإيفالويت/root/yasser/assignment1/assessment2/Assessment 2.0 Paid Flooring Challenge - TAKEOFF-52 Lovesacإنسخ الملفات للمكان الصحيح في ملفات العميل علشان أقدر أعمل عليها إيفالويت
02.05.2026 12:02
ياسر
تزودي subdomain بسرعة؟
02.05.2026 12:09
ياسر
https://opencode.jobotai.site/
02.05.2026 12:34
ياسر
https://opencode.ai/go
02.05.2026 12:47
M
Hey ... Good Morning, Thank you again for the clear message, for the opportunity, and for every chance you’ve given me throughout this process.I completed Assessment 2.0 and submitted everything requested. My initial result was 81.8%. After that, I tested additional heavier algorithms to reduce false positives, even though they require more runtime, but the score dropped to 72%, so I moved quickly with the strongest stable version for submission, because with the current limited resources any extra runs were unlikely to add meaningful improvement. To push performance further, I would need either more time to properly study the available market options and select better-fit algorithms, or access to stronger resources for deeper testing.What I want to highlight is that these results were achieved while using lower resources, fewer tokens, and less execution time. That balance between performance, efficiency, and practical delivery is a core part of how I build systems.This also reflects the confidence I have in my work. My style has always been focused on depth, strong logic, efficient resource usage, and creating systems that perform in real conditions — not just something quick that only works in ideal scenarios.=======================================The repository folder for this submission is here:https://github.com/yasser-saker/assignment1/tree/main/Assignment%202I attached a Loom video inside the repo as well, and here is the direct Loom link too:https://www.loom.com/share/6de5399fb05e4e64bd3035b9cbbaec5a=======================================If you’d like to run or optimize it further, I’d be happy to help. I genuinely believe it can achieve even stronger results on the Mac machine you mentioned.Regarding the point about working with another person, I’d genuinely appreciate discussing this more. My partner and I usually begin projects with brainstorming sessions together (along with AI tools), planning the best structure and direction from the start. In this case, I built the project myself, while he mainly supported early-stage ideation and overall flow while I handled implementation and coding.I want to be fully transparent because I value clarity and directness, and I don’t want to cross any standards or expectations. If there’s room to discuss structure, I’d appreciate being considered as a small agency. We work extremely well together, and I believe that model could create strong value once you review the quality of the work.And yes, if it’s a good fit, we would absolutely be interested in a long-term opportunity with consistent work.Please let me know what your honest thoughts are on the work itself, especially your evaluation of the algorithmic strength and overall approach beyond just the output score. I’d also appreciate knowing what the next step is from your side, whether you’d like revisions, another assessment, an interview, or to discuss onboarding / structure options.
02.05.2026 13:22
ياسر
Weekly retainer
02.05.2026 13:36
ياسر
n8n ?
02.05.2026 14:02
ياسر
فقط مشكلتي مع n8n
02.05.2026 14:03
ياسر
دلوقتي أنا أهم شئ عاوز أنقله هو n8n بكل محتوياته .. إدرس حالته وإكتب تقرير
02.05.2026 14:03
M
https://hpanel.hostinger.com/domain/jobotai.site/dns?tab=dns_records
02.05.2026 15:08
M
mira.z.melhem@gmail.com
02.05.2026 15:09
ياسر
ai.jobotai.site
02.05.2026 15:24
ياسر
https://ai.jobotai.site
02.05.2026 15:28
M
Hey … Good Morning,Thank you again for the clear message, for the opportunity, and for every chance you’ve given me throughout this process.I completed Assessment 2.0 and submitted everything requested. My initial result was 81.8%. After that, I tested additional heavier algorithms to reduce false positives, even though they require more runtime, but the score dropped to 72%, so I moved quickly with the strongest stable version for submission, because with the current limited resources any extra runs were unlikely to add meaningful improvement. To push performance further, I would need either more time to properly study the available market options and select better-fit algorithms, or access to stronger resources for deeper testing.What I want to highlight is that these results were achieved while using lower resources, fewer tokens, and less execution time. That balance between performance, efficiency, and practical delivery is a core part of how I build systems.This also reflects the confidence I have in my work. My style has always been focused on depth, strong logic, efficient resource usage, and creating systems that perform in real conditions — not just something quick that only works in ideal scenarios.=======================================The repository folder for this submission is here:https://github.com/yasser-saker/assignment1/tree/main/Assignment%202I attached a Loom video inside the repo as well, and here is the direct Loom link too:https://www.loom.com/share/6de5399fb05e4e64bd3035b9cbbaec5a=======================================If you’d like to run or optimize it further, I’d be happy to help. I genuinely believe it can achieve even stronger results on the Mac machine you mentioned.Regarding the point about working with another person, I’d genuinely appreciate discussing this more. My partner and I usually begin projects with brainstorming sessions together (along with AI tools), planning the best structure and direction from the start. In this case, I built the project myself, while he mainly supported early-stage ideation and overall flow while I handled implementation and coding.I want to be fully transparent because I value clarity and directness, and I don’t want to cross any standards or expectations. If there’s room to discuss structure, I’d appreciate being considered as a small agency. We work extremely well together, and I believe that model could create strong value once you review the quality of the work.And yes, if it’s a good fit, we would absolutely be interested in a long-term opportunity with consistent work.Please let me know what your honest thoughts are on the work itself, especially your evaluation of the algorithmic strength and overall approach beyond just the output score. I’d also appreciate knowing what the next step is from your side, whether you’d like revisions, another assessment, an interview, or to discuss onboarding / structure options.
02.05.2026 15:43
M
Hi Mira, please accept the contract I sentearlier yesterday with the assessment too. The$50 was already paid for this assessment.Accept the contract inside upwork. Deliver thematerial through there so you can get paid, andI'll keep you posted. Thank you. (edited)
02.05.2026 15:48
ياسر
https://kimi.jobotai.site/https://opencode.jobotai.site/https://ai.jobotai.site/كلهم شغالين عالسيرفر الجديد
02.05.2026 16:12
ياسر
Ollama Cloudcogito-2.1:671bdeepseek-v3.1:671bdeepseek-v3.2deepseek-v4-flashdeepseek-v4-prodevstral-2:123bdevstral-small-2:24bgemini-3-flash-previewgemma3:12bgemma3:27bgemma3:4bgemma4:31bglm-4.6glm-4.7glm-5glm-5.1gpt-oss:120bgpt-oss:20bkimi-k2-thinkingkimi-k2:1tkimi-k2.5kimi-k2.6:cloudminimax-m2minimax-m2.1minimax-m2.5minimax-m2.7ministral-3:14bministral-3:3bministral-3:8bmistral-large-3:675bnemotron-3-nano:30bnemotron-3-superqwen3-coder-nextqwen3-coder:480bqwen3-next:80bqwen3-vl:235bqwen3-vl:235b-instructqwen3.5:397brnj-1:8b
02.05.2026 16:32
ياسر
بيكلف نفس تكلفة gpt-4o-mini
02.05.2026 16:34
ياسر
https://lmstudio.ai/link
02.05.2026 16:38
ياسر
OpenRouterClaude Haiku 3.5Claude Haiku 4.5Claude Opus 4Claude Opus 4.1Claude Opus 4.5Claude Opus 4.6Claude Opus 4.7Claude Sonnet 3.7Claude Sonnet 4Claude Sonnet 4.5Claude Sonnet 4.6Codestral 2508DeepSeek R1 Distill Llama 70BDeepSeek V3 0324DeepSeek V3.1 TerminusDeepSeek V3.1 Terminus (exacto)DeepSeek V3.2DeepSeek V3.2 SpecialeDeepSeek V4 FlashDeepSeek V4 ProDeepSeek-V3.1DeepSeek: R1Devstral 2 2512Devstral MediumDevstral SmallDevstral Small 1.1Elephant (free)FLUX.2 FlexFLUX.2 Klein 4BFLUX.2 MaxFLUX.2 ProFree Models RouterGemini 2.0 FlashGemini 2.5 FlashGemini 2.5 Flash LiteGemini 2.5 Flash Lite Preview 09-25Gemini 2.5 Flash Preview 09-25Gemini 2.5 ProGemini 2.5 Pro Preview 05-06Gemini 2.5 Pro Preview 06-05Gemini 3 Flash PreviewGemini 3 Pro PreviewGemini 3.1 Flash Image Preview (Nano Banana 2)Gemini 3.1 Flash Lite PreviewGemini 3.1 Pro PreviewGemini 3.1 Pro Preview Custom ToolsGemma 2 9BGemma 3 12BGemma 3 12B (free)Gemma 3 27BGemma 3 27B (free)Gemma 3 4BGemma 3 4B (free)Gemma 3n 2B (free)Gemma 3n 4BGemma 3n 4B (free)Gemma 4 26B A4BGemma 4 26B A4B (free)Gemma 4 31BGemma 4 31B (free)GLM 4.5GLM 4.5 AirGLM 4.5 Air (free)GLM 4.5VGLM 4.6GLM 4.6 (exacto)GLM-4.7GLM-4.7-FlashGLM-5GLM-5-TurboGLM-5.1GPT OSS 120BGPT OSS 120B (exacto)GPT OSS 20BGPT OSS Safeguard 20BGPT-4.1GPT-4.1 MiniGPT-4o-miniGPT-5GPT-5 CodexGPT-5 ImageGPT-5 MiniGPT-5 NanoGPT-5 ProGPT-5.1GPT-5.1 ChatGPT-5.1-CodexGPT-5.1-Codex-MaxGPT-5.1-Codex-MiniGPT-5.2GPT-5.2 ChatGPT-5.2 ProGPT-5.2-CodexGPT-5.3-CodexGPT-5.4GPT-5.4 MiniGPT-5.4 NanoGPT-5.4 ProGPT-5.5GPT-5.5 Progpt-oss-120b (free)gpt-oss-20b (free)Grok 3Grok 3 BetaGrok 3 MiniGrok 3 Mini BetaGrok 4Grok 4 FastGrok 4.1 FastGrok 4.20 BetaGrok 4.20 Multi - Agent BetaGrok Code Fast 1Hermes 3 405B Instruct (free)Hermes 4 405BHermes 4 70BIntellect 3Kimi K2Kimi K2 Instruct 0905Kimi K2 Instruct 0905 (exacto)Kimi K2 ThinkingKimi K2.5Kimi K2.6LFM2.5-1.2B-Instruct (free)LFM2.5-1.2B-Thinking (free)Llama 3.2 11B Vision InstructLlama 3.2 3B Instruct (free)Llama 3.3 70B Instruct (free)Mercury 2Mercury Edit 2MiMo-V2-FlashMiMo-V2-OmniMiMo-V2-ProMiMo-V2.5MiMo-V2.5-ProMiniMax M1MiniMax M2MiniMax M2.1MiniMax M2.5MiniMax M2.5 (free)MiniMax M2.7MiniMax-01Mistral Medium 3Mistral Medium 3.1Mistral Small 3.1 24B InstructMistral Small 3.2 24B InstructMistral Small 4Nemotron 3 Nano 30B A3B (free)Nemotron 3 Nano Omni (free)Nemotron 3 SuperNemotron 3 Super (free)Nemotron Nano 12B 2 VL (free)Nemotron Nano 9B V2 (free)nvidia-nemotron-nano-9b-v2o4 MiniPareto Code RouterQwen: Qwen3.5-FlashQwen2.5 Coder 32B InstructQwen2.5 VL 72B InstructQwen3 235B A22B Instruct 2507Qwen3 235B A22B Thinking 2507Qwen3 30B A3B Instruct 2507Qwen3 30B A3B Thinking 2507Qwen3 CoderQwen3 Coder (exacto)Qwen3 Coder 30B A3B InstructQwen3 Coder FlashQwen3 MaxQwen3 Next 80B A3B InstructQwen3 Next 80B A3B ThinkingQwen3.5 397B A17BQwen3.5 Plus 2026-02-15Qwen3.6 PlusRiverflow V2 Fast PreviewRiverflow V2 Max PreviewRiverflow V2 Standard PreviewSeedream 4.5Step 3.5 FlashTrinity Large PreviewTrinity Large ThinkingUncensored (free)
02.05.2026 16:49
ياسر
https://opencode.jobotai.site/
02.05.2026 17:02
ياسر
free models router
02.05.2026 17:12
ياسر
providers and models
02.05.2026 17:16
ياسر
https://artificialanalysis.ai/models/comparisons/deepseek-v4-flash-high-vs-claude-opus-4-5
02.05.2026 18:11
ياسر
هنشتري عند https://openrouter.ai/
02.05.2026 20:28
ياسر
$0.14/M input tokens$0.28/M output tokens
02.05.2026 20:56
ياسر
$0.74/M input tokens$3.49/M output tokens
02.05.2026 20:56
ياسر
opencodejobotai123
02.05.2026 20:58
ياسر
بدل openai
02.05.2026 21:11
ياسر
openai
02.05.2026 21:11
ياسر
نرجع ل openai
02.05.2026 21:11
ياسر
https://amazon.jobotai.site/
02.05.2026 21:20
ياسر
علشان أعمل عليهم RAG
02.05.2026 21:44
ياسر
اقدر استعمل gpt-4o-mini
02.05.2026 21:48
ياسر
opencodejobotai123
02.05.2026 22:06
ياسر
بدل openai
03.05.2026 06:42
ياسر
من هنا ورايح هتبطلي تشحني في openai
03.05.2026 06:43
M
خلينا ندرس ai
03.05.2026 06:45
ياسر
ويصير أوبن راوتر للمشاريع بدل openai
03.05.2026 06:56
ياسر
https://e3asha.jobotai.site/
03.05.2026 06:57
ياسر
🔑 Meta WhatsApp Token📞 Phone Number ID🔐 Webhook Verify Token
03.05.2026 06:59
ياسر
https://e3asha.jobotai.site/
03.05.2026 08:02