rmusser01 · rmusser01 · Nov 28, 2024 · Nov 23, 2024 · Nov 24, 2024 · Nov 26, 2024
diff --git a/App_Function_Libraries/Gradio_UI/Explain_summarize_tab.py b/App_Function_Libraries/Gradio_UI/Explain_summarize_tab.py
diff --git a/App_Function_Libraries/Gradio_UI/Re_summarize_tab.py b/App_Function_Libraries/Gradio_UI/Re_summarize_tab.py
@@ -269,7 +269,7 @@ def resummarize_content(selected_item, item_mapping, content, api_name, api_key=
 
     # Use default prompt if not provided
     if not summarization_prompt:
-        summarization_prompt = config.get('Prompts', 'default_summary_prompt', fallback="""<s>You are a bulleted notes specialist. [INST]```When creating comprehensive bulleted notes, you should follow these guidelines: Use multiple headings based on the referenced topics, not categories like quotes or terms. Headings should be surrounded by bold formatting and not be listed as bullet points themselves. Leave no space between headings and their corresponding list items underneath. Important terms within the content should be emphasized by setting them in bold font. Any text that ends with a colon should also be bolded. Before submitting your response, review the instructions, and make any corrections necessary to adhered to the specified format. Do not reference these instructions within the notes.``` \nBased on the content between backticks create comprehensive bulleted notes.[/INST]
+        summarization_prompt = config.get('Prompts', 'default_summary_prompt', fallback="""<s>You are a bulleted notes specialist. ```When creating comprehensive bulleted notes, you should follow these guidelines: Use multiple headings based on the referenced topics, not categories like quotes or terms. Headings should be surrounded by bold formatting and not be listed as bullet points themselves. Leave no space between headings and their corresponding list items underneath. Important terms within the content should be emphasized by setting them in bold font. Any text that ends with a colon should also be bolded. Before submitting your response, review the instructions, and make any corrections necessary to adhered to the specified format. Do not reference these instructions within the notes.``` \nBased on the content between backticks create comprehensive bulleted notes.
 **Bulleted Note Creation Guidelines**
 
 **Headings**:
@@ -284,7 +284,7 @@ def resummarize_content(selected_item, item_mapping, content, api_name, api_key=
 
 **Review**:
 - Ensure adherence to specified format
-- Do not reference these instructions in your response.</s>[INST] {{ .Prompt }} [/INST]""")
+- Do not reference these instructions in your response.</s>[INST] {{ .Prompt }}""")
 
     # Summarization logic
     summaries = []

diff --git a/App_Function_Libraries/LLM_API_Calls.py b/App_Function_Libraries/LLM_API_Calls.py
diff --git a/App_Function_Libraries/LLM_API_Calls_Local.py b/App_Function_Libraries/LLM_API_Calls_Local.py
diff --git a/App_Function_Libraries/Summarization/Local_Summarization_Lib.py b/App_Function_Libraries/Summarization/Local_Summarization_Lib.py
diff --git a/App_Function_Libraries/Summarization/Summarization_General_Lib.py b/App_Function_Libraries/Summarization/Summarization_General_Lib.py
diff --git a/App_Function_Libraries/Utils/Utils.py b/App_Function_Libraries/Utils/Utils.py
@@ -231,6 +231,7 @@ def load_and_log_configs():
 
         # Local-Models
         kobold_api_ip = config.get('Local-API', 'kobold_api_IP', fallback='http://127.0.0.1:5000/api/v1/generate')
+        kobold_openai_api_IP = config.get('Local-API', 'kobold_openai_api_IP', fallback='http://127.0.0.1:5001/v1/chat/completions')
         kobold_api_key = config.get('Local-API', 'kobold_api_key', fallback='')
 
         llama_api_IP = config.get('Local-API', 'llama_api_IP', fallback='http://127.0.0.1:8080/v1/chat/completions')
@@ -335,6 +336,7 @@ def load_and_log_configs():
             },
             'local_api_ip': {
                 'kobold': kobold_api_ip,
+                'kobold_openai': kobold_openai_api_IP,
                 'llama': llama_api_IP,
                 'ooba': ooba_api_IP,
                 'tabby': tabby_api_IP,

diff --git a/Config_Files/Backup_Config.txt b/Config_Files/Backup_Config.txt
@@ -7,22 +7,23 @@ groq_api_key = <groq_api_key>
 groq_model = llama3-70b-8192
 openai_api_key = <openai_api_key>
 openai_model = gpt-4o
-huggingface_api_key = <huggingface_api_token>
-huggingface_model = CohereForAI/c4ai-command-r-plus
+huggingface_api_key = <huggingface_api_key>
+huggingface_model = meta-llama/Llama-3.1-70B-Instruct
 openrouter_api_key = <openrouter_api_key>
 openrouter_model = mistralai/mistral-7b-instruct:free
-deepseek_api_key = <deepseek-api-key>
+deepseek_api_key = <deepseek_api_key>
 deepseek_model = deepseek-chat
-mistral_model = mistral-large-latest
 mistral_api_key = <mistral_api_key>
-google_api_key =
+mistral_model = mistral-large-latest
+google_api_key = <google_api_key>
 google_model = gemini-1.5-pro
 custom_openai_api_key = <key_here>
 custom_openai_api_ip = <api_ip_here>
 default_api = openai
 
 [Local-API]
 kobold_api_IP = http://127.0.0.1:5001/api/v1/generate
+kobold_openai_api_IP = http://127.0.0.1:5001/v1/chat/completions
 kobold_api_key =
 llama_api_IP = http://127.0.0.1:8080/completion
 llama_api_key =
@@ -37,6 +38,7 @@ ollama_api_key =
 ollama_model = llama3
 aphrodite_api_IP = http://127.0.0.1:8080/completion
 aphrodite_api_key =
+max_tokens = 4096
 
 [Processing]
 processing_choice = cuda
@@ -47,11 +49,19 @@ words_per_second = 3
 save_character_chats = False
 save_rag_chats = False
 local_api_timeout = 90
+local_api_retries = 3
+local_api_retry_delay = 5
+streaming = True
+temperature = 0.7
+top_p = 0.9
+min_p = 0.05
+# https://artefact2.github.io/llm-sampling/
 
 [Auto-Save]
 save_character_chats = False
 save_rag_chats = False
 
+
 [Prompts]
 prompt_sample = "What is the meaning of life?"
 video_summarize_prompt = "Above is the transcript of a video. Please read through the transcript carefully. Identify the main topics that are discussed over the course of the transcript. Then, summarize the key points about each main topic in bullet points. The bullet points should cover the key information conveyed about each topic in the video, but should be much shorter than the full transcript. Please output your bullet point summary inside <bulletpoints> tags. Do not repeat yourself while writing the summary."
@@ -144,6 +154,4 @@ backup_count = 5
 #   gemini-1.5-flash
 #   gemini-1.5-flash-8b
 #   aqa
-#   text-embedding-004
-
-
+#   text-embedding-004
diff --git a/Config_Files/config.txt b/Config_Files/config.txt
@@ -7,22 +7,23 @@ groq_api_key = <groq_api_key>
 groq_model = llama3-70b-8192
 openai_api_key = <openai_api_key>
 openai_model = gpt-4o
-huggingface_api_key = <huggingface_api_token>
-huggingface_model = CohereForAI/c4ai-command-r-plus
+huggingface_api_key = <huggingface_api_key>
+huggingface_model = meta-llama/Llama-3.1-70B-Instruct
 openrouter_api_key = <openrouter_api_key>
 openrouter_model = mistralai/mistral-7b-instruct:free
-deepseek_api_key = <deepseek-api-key>
+deepseek_api_key = <deepseek_api_key>
 deepseek_model = deepseek-chat
-mistral_model = mistral-large-latest
 mistral_api_key = <mistral_api_key>
-google_api_key =
+mistral_model = mistral-large-latest
+google_api_key = <google_api_key>
 google_model = gemini-1.5-pro
 custom_openai_api_key = <key_here>
 custom_openai_api_ip = <api_ip_here>
 default_api = openai
 
 [Local-API]
 kobold_api_IP = http://127.0.0.1:5001/api/v1/generate
+kobold_openai_api_IP = http://127.0.0.1:5001/v1/chat/completions
 kobold_api_key =
 llama_api_IP = http://127.0.0.1:8080/completion
 llama_api_key =
@@ -52,6 +53,7 @@ local_api_timeout = 90
 save_character_chats = False
 save_rag_chats = False
 
+
 [Prompts]
 prompt_sample = "What is the meaning of life?"
 video_summarize_prompt = "Above is the transcript of a video. Please read through the transcript carefully. Identify the main topics that are discussed over the course of the transcript. Then, summarize the key points about each main topic in bullet points. The bullet points should cover the key information conveyed about each topic in the video, but should be much shorter than the full transcript. Please output your bullet point summary inside <bulletpoints> tags. Do not repeat yourself while writing the summary."
@@ -103,13 +105,29 @@ backup_count = 5
 
 #[Comments]
 #OpenAI Models:
-#    f
+#    gpt-4o
+#    gpt-4o-2024-08-06
+#    gpt-4o-mini
+#    o1-preview
+#    o1-mini
+#    text-embedding-3-large
+#    text-embedding-3-small
+#
 #Anthropic Models:
-#    f
+#    claude-3-5-sonnet-20241022
+#    claude-3-5-sonnet-20240620
+#    claude-3-5-haiku-20241022
+#    claude-3-opus-20240229
+#
 #Cohere Models:
-#    f
+#    command-r-plus-08-2024
+#    command-r-plus-04-2024
+#    command-r-08-2024
+#    command-r-03-2024
+#
 #DeepSeek Models:
-#    f
+#    deepseek-chat
+#
 #Groq Models:
 #    f
 #Mistral Models:
@@ -123,9 +141,9 @@ backup_count = 5
 #    open-codestral-mamba
 # Google's Models (11/15/2024): https://ai.google.dev/gemini-api/docs/models/gemini
 #   gemini-1.5-pro
+#   gemini-1.5-pro-2
+#   LearnLM
 #   gemini-1.5-flash
 #   gemini-1.5-flash-8b
 #   aqa
-#   text-embedding-004
-
-
+#   text-embedding-004
diff --git a/Docs/Bitnet_and_Mamba_Papers.md b/Docs/Bitnet_and_Mamba_Papers.md
@@ -0,0 +1,18 @@
+# Bitnet and MAMBA Papers
+
+
+
+### Bitnet
+- Unsorted
+  - https://arxiv.org/abs/2411.11843
+
+
+
+### Mamba
+- Unsorted
+  - f
+  - f
+  - f
+
+
+
diff --git a/Docs/Code_Documentation/Database.md b/Docs/Code_Documentation/Database.md
@@ -9,6 +9,7 @@
 
 --------------------
 ### SQLite
+https://mrsuh.com/articles/2024/sqlite-index-visualization-structure/
 - **101**
     * https://www.sqlite.org/draft/about.html
 - **Concurrency**

diff --git a/Docs/Design/DB_Design.md b/Docs/Design/DB_Design.md
@@ -0,0 +1,45 @@
+# DB Design
+
+## Table of Contents
+- [General DB](#general-db)
+- [SQLite](#sqlite)
+- [SQLite DB Design](#sqlite-db-design)
+- [Interesting/Relevant Later](#interesting-relevant-later)
+
+https://kerkour.com/sqlite-for-servers
+https://wafris.org/blog/rearchitecting-for-sqlite
+    General DB:
+    https://en.wikipedia.org/wiki/Database_normalization
+    https://www.kimballgroup.com/data-warehouse-business-intelligence-resources/kimball-techniques/dimensional-modeling-techniques/
+
+    Gradio:
+    https://www.gradio.app/guides/connecting-to-a-database
+    https://www.gradio.app/docs/gradio/dataset
+
+    SQLite:
+    https://www.sqlite.org/queryplanner.html
+    https://www.sqlite.org/optoverview.html
+    https://www.sqlite.org/queryplanner-ng.html
+    https://docs.python.org/3/library/sqlite3.html
+
+    SQLite Vector Search:
+    https://github.com/asg017/sqlite-vec
+    https://alexgarcia.xyz/blog/2024/building-new-vector-search-sqlite/index.html
+    https://news.ycombinator.com/item?id=40243168
+    [test-tokenizer-random.py](..%2F..%2F..%2FGithub%2Fllama.cpp%2Ftests%2Ftest-tokenizer-random.py)
+    SQLite DB Design:
+    https://stackoverflow.com/questions/66293837/smart-way-to-structure-my-sqlite-database?rq=3
+    https://stackoverflow.com/questions/7235435/sqlite-structure-advice?rq=3
+    https://stackoverflow.com/questions/19368506/very-basic-sqlite-table-design?rq=3
+    https://stackoverflow.com/questions/7665735/how-do-i-organize-such-database-in-sqlite?rq=3
+    https://stackoverflow.com/questions/29055263/sql-database-layout-design?rq=3
+
+    Seems like it might be interesting/relevant later:
+    https://docs.llamaindex.ai/en/stable/examples/index_structs/struct_indices/SQLIndexDemo/
+
+
+    https://www.microsoft.com/en-us/research/blog/graphrag-unlocking-llm-discovery-on-narrative-private-data/
+    https://www.youtube.com/watch?v=r09tJfON6kE
+
+
+    error handling, vacuuming, zipping transcriptions above X size, and external storage for documents and said zipped transcriptions
diff --git a/Docs/Design/Ideas.md b/Docs/Design/Ideas.md
@@ -1,5 +1,5 @@
 
-https://wafris.org/blog/rearchitecting-for-sqlite
+
 
 Multi-modal support
 Collaborative features: Consider adding features that allow multiple users to collaborate on content analysis, annotation, or editing. This could include shared workspaces or real-time collaborative editing.

diff --git a/Docs/Handy_Dandy_Papers.md b/Docs/Handy_Dandy_Papers.md
@@ -0,0 +1,36 @@
+# Handy Dandy Papers
+
+
+### Context
+    https://arxiv.org/pdf/2304.12102
+    https://arxiv.org/html/2405.20234v2#S3
+
+### Reasoning
+- https://huggingface.co/papers/2411.15862
+
+### Test-Time Compute
+- https://github.com/GAIR-NLP/O1-Journey
+
+
+### Personalization
+    https://arxiv.org/abs/2411.16034
+
+
+### RAG
+    https://arxiv.org/html/2407.21059v1
+
+### Sampling
+    https://arxiv.org/abs/2411.09661
+
+
+### Storytelling
+    https://dreamrunner-story2video.github.io/
+
+
+### Structured Generation
+    https://arxiv.org/abs/2411.15100
+
+### Test-Time related
+    https://mathcritique.github.io/
+
+
diff --git a/Docs/Citations_and_Confabulations.md → Docs/Issues/Citations_and_Confabulations.md b/Docs/Citations_and_Confabulations.md → Docs/Issues/Citations_and_Confabulations.md
@@ -19,6 +19,17 @@ Finetuning:
     - https://mattyyeung.github.io/deterministic-quoting#7-conclusion-is-this-really-ready-for-healthcare
 
 
+
+Anthropic:
+```
+Draft a press release for our new cybersecurity product, AcmeSecurity Pro, using only information from these product briefs and market reports.
+<documents>
+{{DOCUMENTS}}
+</documents>
+
+After drafting, review each claim in your press release. For each claim, find a direct quote from the documents that supports it. If you can’t find a supporting quote for a claim, remove that claim from the press release and mark where it was removed with empty [] brackets.
+```
+
 Attributions
   https://github.com/aws-samples/llm-based-advanced-summarization/blob/main/detect_attribution.ipynb
 
@@ -75,12 +86,17 @@ Research
   https://www.lycee.ai/blog/rag-ragallucinations-and-how-to-fight-them
   https://arxiv.org/abs/2407.16557
   https://arxiv.org/abs/2407.16604
+  https://arxiv.org/pdf/2309.11495
+  https://deepmind.google/research/publications/85420/
   https://thetechoasis.beehiiv.com/p/eliminating-hallucinations-robots-imitate-us
   https://arxiv.org/abs/2407.19825
+  https://arxiv.org/abs/2411.14257
   https://arxiv.org/pdf/2406.02543
   https://arxiv.org/abs/2406.10279
+  https://arxiv.org/abs/2402.17811
   https://arxiv.org/pdf/2409.18475
   https://llm-editing.github.io/
+  https://arxiv.org/abs/2411.14257
   https://arxiv.org/pdf/2407.03651
   https://cleanlab.ai/blog/trustworthy-language-model/
   https://arxiv.org/abs/2408.07852