You need to agree to share your contact information to access this model
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
PredictRAM Vedaniq is provided under controlled research access. Please submit your details for manual review. Access may be granted for research, evaluation, internal experimentation, or approved commercial discussion only. This model does not provide investment advice and must not be used as the sole basis for trading, investing, compliance, or financial decisions.
Log in or Sign Up to review the conditions and access this model content.
PredictRAM Vedaniq v1
PredictRAM Vedaniq is a finance-first language model that converts a single prompt into analyst-grade research, valuation, derivatives logic, portfolio insight, and financial modeling workflows.
Designed for analysts, investors, finance teams, and market professionals, Vedaniq focuses on Indian capital markets and structured financial reasoning rather than generic chatbot behavior.
What It Can Do
- Generate algo model ideas with 1Prompt
- Review financial reports with 1Prompt
- Assist in connecting trading logic with broker or trading APIs with 1Prompt
- Prepare DCF valuation model structures with 1Prompt
- Support complex market simulation workflows with 1Prompt
- Assist with leveraged buyout (LBO) model building with 1Prompt
- Support merger and acquisition (M&A) analysis workflows with 1Prompt
- Assist with complex derivatives and options pricing model design with 1Prompt
- Help structure balance sheet optimization models with 1Prompt
Core Capabilities
- Equity research and stock analysis
- Valuation framing and financial modeling support
- Options and derivatives strategy explanation
- Portfolio concentration and risk commentary
- Financial education support for CFA, FRM, MBA, and analyst training
- Corporate finance and strategic analysis workflows
- Research automation and structured finance output generation
Positioning
Vedaniq is built as a financial co-pilot for high-value analytical work. It is optimized for structured outputs, finance-first reasoning, and single-prompt productivity across research, valuation, risk, and strategy.
Current Benchmark Snapshot
- Overall composite score: 85.3 / 100
- Phase 1 structure score: 95.0 / 100
- Phase 2 structure score: 60.0 / 100
- Overall pass rate: 7 / 9
- Known limitations: slower inference latency and weaker consistency on selected options and education prompts
Intended Use
This model is intended for:
- Research and evaluation
- Internal experimentation
- Analyst productivity workflows
- Financial education and training support
- Approved enterprise and professional review use
Not Intended For
This model is not intended to be used as:
- Investment advice
- Trading advice
- Legal advice
- Tax advice
- Regulatory advice
- A fully autonomous execution or decision engine without human review
Access Notice
This model is distributed under controlled gated access. By requesting access, you acknowledge and agree that:
- The model is provided for research, evaluation, and approved internal use only.
- The model output does not constitute investment advice, research advice, legal advice, or regulatory advice.
- You will independently validate all outputs before any business, investment, trading, or compliance use.
- You will not misrepresent model outputs as certified financial recommendations.
- Access may be revoked at any time at the discretion of the model owner.
Legal Disclaimer
PredictRAM Vedaniq may generate incomplete, outdated, or incorrect information. Market data, valuations, option strategy commentary, macro interpretation, modeling logic, and educational content may contain errors or omissions. Users are solely responsible for verification, suitability assessment, regulatory compliance, and downstream decisions. PredictRAM and its affiliates disclaim liability for any loss, damage, claim, or regulatory consequence arising from use of this model.
Technical Notes
- Base model: Qwen/Qwen2.5-7B-Instruct
- Fine-tuning approach: QLoRA
- Primary focus: Indian capital markets and finance workflows
- Delivery mode: controlled gated access
Contact
For enterprise access, partnership discussions, or controlled evaluation requests, use the gated access request form on this model page.
PredictRAM Vedaniq — Financial Language Model
A QLoRA fine-tune of Qwen/Qwen2.5-7B-Instruct specialised for Indian capital markets (NSE/BSE equities, F&O, SEBI/RBI context, valuation, portfolio analysis, macro interpretation).
Capabilities
| Mode | Description |
|---|---|
| Equity analyst | Single-stock grounded research with valuation, risks, price action |
| Portfolio copilot | Multi-ticker concentration, sector, risk commentary |
| Options explainer | Strategy explanation with Greeks + NSE context |
| Education copilot | CFA/FRM/MBA finance prep |
Quick start
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
model_id = "isubirsingh/predictram-vedaniq-v1"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.float16,
device_map="auto",
)
system = (
"You are PredictRAM's Financial SLM for Indian capital markets. "
"Provide concise, data-driven answers for NSE/BSE stocks, "
"macroeconomics, SEBI context, valuation, risk, and portfolio analysis."
)
prompt = "Analyse the working capital cycle for an Indian FMCG company."
messages = [
{"role": "system", "content": system},
{"role": "user", "content": prompt},
]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=512)
print(tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))
Training details
- Base model: Qwen/Qwen2.5-7B-Instruct
- Method: QLoRA (4-bit NF4, r=16, alpha=32)
- Dataset: Proprietary PredictRAM financial Q&A dataset (Indian market focus)
- Hardware: AWS EC2 g5.12xlarge (A10G 24 GB)
Intended use & limitations
This model is for research and evaluation only and should not be used as the sole basis for investment decisions. Financial data cited in outputs may be outdated. Always verify figures with official regulatory filings.
Citation
@misc{predictram-vedaniq-2025,
title = {PredictRAM Vedaniq: An Indian Capital Markets Language Model},
author = {PredictRAM},
year = {2025},
url = {https://huggingface.co/isubirsingh/predictram-vedaniq-v1}
}
- Downloads last month
- -