R0 BLSP cohort: 5 frozen-frozen adapter checkpoints

.gitattributes

@@ -0,0 +1,3 @@
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text

A_BLSP_seed1234.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5cc55d352a3bcd7a147896f49286bae6f2719465d71174c161673377b2fb2ee
+size 335649312

A_BLSP_seed1235.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6342f7b6e2406ced8a887917f8a5ffa7c1b4aa57ed7d76f16df94f25e035e6d2
+size 335648160

A_BLSP_seed1236.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ebd1ddd9d0abc197436741c91d097ba431eb8d84ba1b47d4e61f88dfba9368a
+size 335648160

A_BLSP_seed1237.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d31deafaa33e1f33bcd9f959de8d14c64f403bc6153e30267eb8c60426b593a6
+size 335648160

A_BLSP_seed1238.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d25ca9da1fe215bc06523fa40ee013804786a79a24a1003d827239dbdfce8a3a
+size 335648160

README.md

@@ -0,0 +1,64 @@
+---
+license: cc-by-nc-4.0
+tags:
+  - speech-language-models
+  - sentence-stress
+  - frozen-frozen
+  - blsp
+  - paralinguistic
+base_model: Qwen/Qwen3-8B
+---
+
+# BTA — R0 BLSP cohort (5 seeds)
+
+5-seed cohort of the BLSP-class baseline adapter for the
+Beyond Transcript Alignment research project.
+
+**Architecture:** Conv1d(1024→4096, k=4, s=4) → MLP-2 → RMSNorm → +modality token (84M trainable params)
+**Encoder:** WavLM-Large layer 16 (`microsoft/wavlm-large`, frozen)
+**LLM:** Qwen3-8B-Instruct (`Qwen/Qwen3-8B`, frozen, `enable_thinking=False`)
+**Loss:** $\mathcal{L}_{\mathrm{BLSP}} = \mathcal{L}_{\mathrm{task}} + \lambda_{\mathrm{KL}} \mathcal{L}_{\mathrm{KL}}$
+**Optimizer:** AdamW lr=5e-5, weight_decay=0.01, max_norm=1.0; cosine to 1e-6
+**Schedule:** seed 1234 = 600 steps + warmup 500; seeds 1235-1238 = 300 steps + warmup 250
+**Code:** <https://github.com/Nurgali-Kadyrbek/frozen-speech-llm-stress>
+
+## Files
+
+| File | Size |
+|---|---|
+| `A_BLSP_seed1234.pt` | 336 MB |
+| `A_BLSP_seed1235.pt` | 336 MB |
+| `A_BLSP_seed1236.pt` | 336 MB |
+| `A_BLSP_seed1237.pt` | 336 MB |
+| `A_BLSP_seed1238.pt` | 336 MB |
+| `adapter_init.json` | calibrated init (std_8B=0.02205, RMSNorm scale=0.0215) |
+
+## Reported metrics (5-seed cohort)
+
+| Metric | Mean ± σ |
+|---|---|
+| Probe-G total | 0.6306 ± 0.0095 |
+| **Probe-G$_{\mathrm{neutral}}$** | **0.5122 ± 0.0108** |
+| Probe-G$_{\mathrm{explicit}}$ | 0.7491 ± 0.0120 |
+| Probe-K linear eval_full | 0.2105 ± 0.0128 |
+| Probe-K MLP-2 eval_full | 0.2446 ± 0.0339 |
+
+**Sub-cascade collapse:** every seed's Probe-K linear is below the
+$K_T$ text-only baseline of 0.290.
+
+## Usage
+
+```python
+import torch
+from huggingface_hub import hf_hub_download
+
+ckpt = hf_hub_download("nur-dev/frozen-stress-r0-blsp", "A_BLSP_seed1234.pt")
+state = torch.load(ckpt, map_location="cpu")
+# Adapter loading — see scripts/stage2_eval.py in the GitHub repo
+```
+
+## License
+
+CC-BY-NC-4.0 (weights inherit non-commercial restriction from the
+training datasets Stress-17K-raw, StressPresso, Expresso). Permitted
+for academic research and ablation studies.

adapter_init.json

@@ -0,0 +1,28 @@
+{
+  "qwen3_8b": {
+    "model_id": "Qwen/Qwen3-8B",
+    "d_llm": 4096,
+    "embed_tokens_std": 0.022050147876143456,
+    "embed_row_norm_mean": 1.3758326768875122
+  },
+  "adapter_init": {
+    "d_enc": 1024,
+    "d_llm": 4096,
+    "conv_kernel": 4,
+    "conv_stride": 4,
+    "mlp_hidden_mult": 2,
+    "last_linear_std": 0.0003445335605647415,
+    "rmsnorm_init_scale": 0.021490697426189264,
+    "modality_token_std": 0.022050147876143456
+  },
+  "smoke": {
+    "median_ratio": 1.4409220218658447,
+    "p10_ratio": 1.4286571741104126,
+    "p90_ratio": 1.4487676620483398,
+    "target_band": [
+      0.5,
+      2.0
+    ],
+    "n_trainable_params": 83910656
+  }
+}