Upload README.md with huggingface_hub

README.md

@@ -7,103 +7,23 @@ tags:
 - salesforce
 - apex
 - lwc
-- lightning-web-components
 - code
 - fine-tuned
 - gguf
-- unsloth
 pipeline_tag: text-generation
 ---
 
 # DeepForce Coder v1
 
-A Salesforce-specialized AI coding assistant fine-tuned on Qwen 2.5 Coder 3B.
-Built specifically for Salesforce developers to generate, debug, review, and refactor
-Apex code and Lightning Web Components following enterprise best practices.
+> ⚠️ **Note:** v1 has known issues with simple Apex generation due to 
+> adapter weight loss during training. v2 is currently in development 
+> with full retraining and will be significantly improved.
+> Use v2 when available: coming soon.
 
-## Capabilities
+## Known Limitations in v1
+- Over-engineers simple Apex requests
+- Occasionally generates non-existent Apex APIs
+- Best used for complex generation, debug, review, and refactor tasks
 
-| Task | Description |
-|------|-------------|
-| Apex Generation | Write production-ready Apex classes, triggers, batch, scheduled, queueable |
-| LWC Development | Create Lightning Web Components with SLDS conventions |
-| Code Debug | Identify bugs with severity ratings and corrections |
-| Code Review | Review code against Salesforce best practices |
-| Refactoring | Simplify over-engineered code while preserving security |
-| Test Classes | Generate comprehensive Apex test classes |
-
-## Best Practices Enforced
-
-- `with sharing` on all classes
-- `WITH USER_MODE` on all SOQL queries
-- `Security.stripInaccessible()` before DML
-- `try-catch` on all DML and callouts
-- `Database.update/insert(records, false)` for bulk DML
-- No SOQL or DML inside loops
-- Bulkified trigger handlers with recursion guards
-
-## Model Details
-
-- **Base model**: Qwen/Qwen2.5-Coder-3B-Instruct
-- **Fine-tuning**: LoRA adapters across 8 specialized Salesforce tasks
-- **Training data**: curated Salesforce-specific examples
-- **Quantization**: Q4_K_M GGUF (1.80 GB)
-- **Context length**: 6144 tokens
-
-## Quick Start
-
-### Ollama
-```bash
-ollama run hf.co/deepforce/deepforce-coder-v1:Q4_K_M
-```
-
-### llama.cpp
-```bash
-llama serve -hf deepforce/deepforce-coder-v1:Q4_K_M
-```
-
-### Python (llama-cpp-python)
-```python
-from llama_cpp import Llama
-
-llm = Llama.from_pretrained(
-    repo_id  = "deepforce/deepforce-coder-v1",
-    filename = "deepforce-coder-v1-q4_k_m.gguf",
-)
-
-response = llm.create_chat_completion(messages=[
-    {"role": "system", "content": "You are DeepForce Coder, an expert Salesforce developer."},
-    {"role": "user",   "content": "Write a simple Apex class that returns Accounts by industry."}
-])
-print(response["choices"][0]["message"]["content"])
-```
-
-## Example Prompts
-
-**Generate Apex:**
-
-Write a trigger handler for Opportunity that creates a follow-up Task when StageName changes to Closed Won.
-
-**Debug Apex:**
-Debug the following Apex code: [paste your code]
-
-**Review Apex:**
-Review the following Apex code for best practices: [paste your code]
-
-**Generate LWC:**
-Create an LWC component that displays a list of Accounts in a lightning-datatable.
-
-## Limitations
-
-- v1 release — some outputs may occasionally use Java syntax patterns
-- Test class generation uses System.assertEquals instead of Assert class in some cases
-- These will be fixed in v2
-
-## Training
-
-Fine-tuned using [Unsloth](https://github.com/unslothai/unsloth) on Google Colab.
-Training data generated using Anthropic Claude API.
-
-## License
-
-Apache 2.0 — free for commercial and personal use.
+## v2 Coming Soon
+Full retraining with verified adapter weights across all adapters.