SAVRN Model Hub · Comparisons
bge-base-en-v1.5-course-recommender-v5 vs pubmedbert-base-embeddings
Bge-base-en-v1.5-course-recommender-v5 has 109M parameters and pubmedbert-base-embeddings has 109M parameters; at 16-bit, bge-base-en-v1.5-course-recommender-v5 needs about 0.3 GB (1x MI300X from $1.85 an hour) and pubmedbert-base-embeddings about 0.3 GB (1x MI300X from $1.85 an hour).
| Field | bge-base-en-v1.5-course-recommender-v5 datasocietyco/bge-base-en-v1.5-course-recommender-v5 | pubmedbert-base-embeddings NeuML/pubmedbert-base-embeddings |
|---|---|---|
| Publisher | Data Society | NeuML |
| Task | Sentence similarity | Sentence similarity |
| Modality | Text | Text |
| Parameters, as reported | 109M parameters | 109M parameters |
| Architecture | BertModel | BertModel |
| Library | sentence-transformers | sentence-transformers |
| Context length | 512 tokens | 512 tokens |
| Repository size | 438.9 MB | 876.9 MB |
| Artifact formats | safetensors | safetensors, pytorch |
| License | Not stated | apache-2.0 |
| Access | Open weights, no gate | Open weights, no gate |
| Memory at 16-bit (weights and margin) | 0.3 GB | 0.3 GB |
| Cheapest GPUs at 16-bit, per hour | 1x MI300X, $1.85 | 1x MI300X, $1.85 |
| Memory at 4-bit (weights and margin) | 0.1 GB | 0.1 GB |
| Cheapest GPUs at 4-bit, per hour | 1x MI300X, $1.85 | 1x MI300X, $1.85 |
| Revision viewed | 2b069eed51ce | b79526d6ef36 |
| Downloads reported by the hub | 4.1M | 1.1M |
| Last observed | 2026-10-10 | 2026-10-10 |
An evaluation row appears only where at least two of these models report the same benchmark with the same stated configuration, metric, unit and setup. Different evaluators stay named in each cell. Values are shown as reported: no unit conversion, no ranking.
SAVRN's Notes on bge-base-en-v1.5-course-recommender-v5
The name is the use case: Data Society's fine-tune of BAAI/bge-base-en-v1.5 aimed at course recommendation, turning each input into a 768-dimensional vector for similarity search. The arithmetic is trivial — 109M parameters, 0.3 GB at 16-bit, 0.1 GB at 4-bit, a 438 MB download — so the one MI300X the Index lists at $1.85 an hour is wasted on it alone; run it beside your serving stack. The 512-token context still applies, so long course descriptions get chunked before embedding.
Treat the license as the gating item: the record lists none for this fine-tune, which makes commercial use a conversation with the publisher, not a checkbox. The disclosed training is a light touch — 24 steps at batch size 16, learning rate 3e-06 — so validate the delta against the base on your own catalog before shipping; a fine-tune this small can be a wash. Published January 2025 under sentence-transformers, derived from bge-base-en-v1.5, and drawing 4.1M downloads a month.
SAVRN's Notes on pubmedbert-base-embeddings
Medical literature search is where we would put this one. NeuML fine-tuned a PubMedBERT base with sentence-transformers on PubMed title and abstract pairs, and the output is a 768-dimensional vector per sentence or paragraph, built for clustering and semantic search over medical text. The 109M parameters need 0.3 GB at 16-bit, and the stored precision is float32, so even the full-width copy stays under 1 GB. The lowest-cost host in our data is one MI300X at $1.85 an hour, far more card than an embedding job this size needs. Share it with something bigger.
Apache 2.0 covers commercial use, modification and redistribution, with notices kept and significant changes stated. Two things to check: the 512-token context, so anything longer gets chunked before embedding, and the parent model, Microsoft's BiomedBERT base trained on abstracts and full text, which carries its own terms.
Questions
Which is larger, bge-base-en-v1.5-course-recommender-v5 or pubmedbert-base-embeddings?
bge-base-en-v1.5-course-recommender-v5 (109M parameters) is larger than pubmedbert-base-embeddings (109M parameters), by the parameter counts their publishers report.
Which is cheaper to run, bge-base-en-v1.5-course-recommender-v5 or pubmedbert-base-embeddings?
At 4-bit, bge-base-en-v1.5-course-recommender-v5 fits on 1x MI300X from $1.85 an hour and pubmedbert-base-embeddings on 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use pubmedbert-base-embeddings commercially?
Yes. pubmedbert-base-embeddings is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.